Hastimal Jangid
Co-Founder, RankRabbit.ai | Coozmoo - Cloud and AI Engineering
Houston, Texas, United States
Actions
Hastimal Jangid is a Cloud, Data, and AI technology leader, enterprise architect, researcher, and technical speaker with 15+ years of experience designing large-scale distributed, data-intensive, and AI-powered systems. He is 10× AWS, 7× GCP, 3×Azure, Kubernetes, Snowflake, Databricks, GenAI, GitHub, and Anthropic MCP Certified, combining deep multi-cloud expertise with emerging AI technologies.
His work and research focus on Large Language Models (LLMs), Agentic AI, AI Search and Retrieval, RAG, Model Context Protocol (MCP), cloud-native AI infrastructure, real-time data engineering, AI security, and responsible AI, with particular interest in applications across healthcare and biomedical systems.
As Co-founder of RankRabbit AI, Hastimal explores how LLMs and AI agents discover, retrieve, rank, reason over, and recommend information across platforms such as Gemini, ChatGPT, Claude, Perplexity, and Google’s AI-powered search experiences. His research interests include LLM retrieval and grounding, multi-agent systems, AI search, information quality, explainable AI (xAI), knowledge-base poisoning, AI security, and trustworthy AI systems.
His engineering background spans AWS, Google Cloud, Microsoft Azure, Kubernetes, distributed systems, Big Data, machine learning, MLOps, real-time streaming architectures, and AI agent infrastructure. He has designed and delivered enterprise technology initiatives across healthcare, financial services, retail, travel, media, and digital commerce.
As a technical speaker, Hastimal translates emerging AI research into practical architectures and hands-on demonstrations. His talks explore the intersection of LLMs, AI agents, MCP, cloud infrastructure, K8, search, security, and healthcare AI—explaining how modern AI systems work, where they fail, and how engineers can build them securely, responsibly, and at scale.
Badges
Area of Expertise
Topics
Beyond Chatbots: Architecting Multi-Agent AI Systems with Microsoft Foundry
Not every AI problem should be solved with one model and one prompt. Complex applications increasingly use specialized agents for planning, research, tool execution, validation, and synthesis. This session explores how to design multi-agent systems using Microsoft Foundry and modern agentic architecture patterns. We will examine agent responsibilities, orchestration, context sharing, tool access, handoffs, failure recovery, evaluation, and human-in-the-loop controls. Through an end-to-end example, attendees will learn when multiple agents improve a system, when they create unnecessary complexity, and how to design agent workflows that remain observable, testable, secure, and maintainable.
AgentOps on Azure: Observability, Evaluation, and Reliability for Production AI Agents
Traditional application monitoring is not enough for systems where an LLM decides what to do next. Production AI agents require visibility into prompts, model responses, retrieval, tool calls, agent handoffs, latency, cost, failures, and evaluation results. This session introduces practical AgentOps patterns for Azure-based agentic applications. We will explore tracing complete agent workflows, evaluating non-deterministic outputs, monitoring tool execution, detecting regressions, measuring grounding quality, and integrating agent evaluation into delivery pipelines. Attendees will learn how to make agent systems observable, testable, and operable so production failures can be diagnosed rather than hidden inside a generated response.
Focus: Operating AI agents in production: tracing, evaluation, tool-call monitoring, reliability, testing, latency, cost, CI/CD, and incident investigation.
Beyond Accuracy: Measuring the Value of Evidence-Grounded AI in Healthcare
Healthcare organizations are rapidly adopting generative and agentic AI, but evaluating these systems requires more than measuring answer accuracy. In healthcare, value also depends on whether AI retrieves authoritative evidence, preserves provenance, supports verification, and reduces the effort required to find trustworthy information.
This session presents a practical framework for evaluating evidence-grounded agentic AI across retrieval relevance, source authority, citation fidelity, provenance completeness, evidence coverage, response traceability, and search-effort reduction. Using a working healthcare discovery prototype, we demonstrate how separating query planning, retrieval, evidence ranking, and grounded response generation makes each stage measurable and auditable. The session explores how healthcare organizations can move beyond model-centric metrics and evaluate whether AI systems create real informational, operational, and clinical-support value.
Trust Your Data Before Your AI: Governance, Lineage, and Security with Fabric and Purview
AI and self-service analytics make enterprise data easier to consume—but they can also amplify bad data, unclear definitions, inappropriate access, and governance gaps. This session explores how Microsoft Fabric and Microsoft Purview can help organizations build a trusted foundation for analytics and AI. We will examine data lineage, discovery, ownership, access controls, classification, trusted data products, and governance across Fabric workloads and Power BI. The session connects governance to an increasingly important question: if an AI assistant or analytics user produces an answer from enterprise data, can we determine where that data came from and whether it should have been used?
Focus: Governance for Fabric and AI-ready analytics: lineage, data ownership, security, trusted datasets, metadata, and controlled access.
Zero-Trust AI Agents: Securing MCP, Tools, Identity, and Permissions on Azure
AI agents are becoming active participants in enterprise systems—calling APIs, accessing data, invoking MCP tools, and taking actions on behalf of users. That creates a new security boundary. This session presents a zero-trust architecture for securing agentic applications on Azure. We will examine agent identity, least-privilege authorization, managed identities, secrets, tool permissions, MCP security, prompt injection, approval gates, and auditability. Through practical architecture patterns and attack scenarios, attendees will learn how to prevent an AI agent from becoming an overly privileged automation layer while still enabling useful autonomous workflows.
Focus: Securing autonomous AI agents, MCP tools, identities, permissions, APIs, secrets, and privileged actions using zero-trust architecture.
Running AI Agents at Scale: Azure Container Apps vs. AKS for Agentic Workloads
Once an AI agent moves beyond a prototype, teams must decide where and how to run it. Should an agentic workload use Azure Container Apps or Azure Kubernetes Service? This session compares both approaches through the lens of production AI agents. We will examine containerization, scaling, networking, workload identity, secrets, observability, resiliency, background workers, tool services, deployment strategies, and cost. Using practical agent architectures, attendees will learn how runtime requirements influence platform selection, how to avoid unnecessary Kubernetes complexity, and when AKS provides the control and extensibility a production system actually requires.
Focus: Choosing and designing the runtime platform for production AI agents using Azure Container Apps or Azure Kubernetes Service.
From Real-Time Data to Action: Event-Driven Analytics with Microsoft Fabric
Dashboards tell us what happened. Modern analytics platforms increasingly need to recognize what is happening now and help teams respond immediately. This session explores an event-driven analytics architecture using Microsoft Fabric, real-time data, Power BI, and Data Activator capabilities. We will follow streaming events from ingestion through analysis and visualization, then demonstrate how business conditions can trigger downstream actions and notifications. The session will also cover architecture decisions around latency, event modeling, operational analytics, alert fatigue, and when real-time processing actually provides value. Attendees will leave with reusable patterns for moving from passive dashboards toward responsive, event-driven analytics.
Focus: Streaming and event-driven analytics: ingesting real-time data, detecting business conditions, visualizing them, and triggering actions.
Objectives:
1. Design a real-time analytics pipeline that moves streaming events into Fabric and Power BI experiences.
2. Define meaningful business conditions and use event-driven patterns to turn analytics signals into actions.
3. Evaluate latency, scale, reliability, and operational trade-offs when deciding whether a workload truly requires real-time analytics.
OneLake to Insight: Designing a Production-Ready Microsoft Fabric Data Platform
A Microsoft Fabric proof of concept can be built quickly, but designing a Fabric platform for production requires decisions about far more than individual workloads. This session presents an end-to-end architecture for moving data from source systems through ingestion, OneLake, data engineering and warehousing, semantic models, and Power BI. We will examine workspace design, data organization, security, governance, performance, workload boundaries, and operational considerations. Rather than treating Fabric services as isolated features, attendees will learn how the pieces fit together as one enterprise analytics platform and how to make practical architecture decisions that remain maintainable as data volume, teams, and use cases grow.
Focus: End-to-end Fabric architecture: ingestion, OneLake, engineering, warehouse/lakehouse, semantic models, Power BI, governance, and operational design.
Objectives:
1. Design an end-to-end Fabric architecture from ingestion and OneLake through analytics and Power BI.
2. Choose appropriately between Fabric Data Engineering, Data Factory, Lakehouse, and Warehouse patterns based on workload requirements.
3. Apply production architecture principles for security, governance, performance, workspace organization, and maintainability.
From Cloud Engineer to AI Architect: Staying Relevant in the Agentic AI Era; A founder's journey
Agentic AI is changing what it means to be a cloud engineer, developer, architect, or DevOps professional. The most valuable engineers will not simply learn another AI tool; they will learn how to combine software, cloud, data, security, and AI into complete systems. This session presents a practical framework for evolving from traditional cloud and application roles toward AI architecture and agentic engineering. We will explore which existing skills become more valuable, which new competencies matter, how to build credible hands-on experience, and how engineers can remain technically relevant without chasing every new model or framework.
Focus: Career transformation for cloud, DevOps, data, and software professionals as AI agents change engineering roles.
From Prototype to Production: Building Enterprise AI Agents on Azure
Building an AI agent demo is easy. Running one reliably in production is not. This session walks through the architecture of a production-ready agentic application on Azure, covering model access, grounding, tool calling, MCP, identity, evaluation, observability, reliability, and deployment. We will examine how Microsoft Foundry and Azure services can support the lifecycle from prototype to enterprise deployment, while addressing latency, cost, security, failure handling, and human oversight. Attendees will leave with a practical reference architecture and decision framework for moving AI agents beyond experimentation and into real-world applications.
Focus: Microsoft Foundry, tool calling, MCP, grounding, evaluation, deployment, observability, reliability, cost, and production readiness.
Beyond Dashboards: Building AI-Ready Analytics with Microsoft Fabric and Power BIq
AI is changing how users interact with analytics, but reliable AI experiences still depend on well-designed data, semantic models, and governance. This session explores how Microsoft Fabric and Power BI can provide the foundation for AI-ready analytics. We will follow data from OneLake and Fabric engineering workloads through semantic modeling and Power BI, then examine how AI-assisted experiences can consume that trusted context. Along the way, we will discuss data quality, metadata, business definitions, security, and why simply connecting an LLM to enterprise data is not enough. Attendees will leave with a practical architecture for preparing analytics platforms for the AI era.
Focus: How Fabric, semantic models, Power BI, and AI work together to turn governed enterprise data into AI-ready analytics.
Objectives:
1. Design an AI-ready analytics architecture connecting Fabric, OneLake, semantic models, and Power BI.
2. Understand how data quality, metadata, business semantics, and governance influence the reliability of AI-generated insights.
3. Identify where AI adds value to analytics workflows and where traditional BI remains the better approach.
Beyond RAG: Designing the Data Layer for Enterprise AI Agents on Azure
Enterprise AI agents are only as reliable as the data they can retrieve. Moving beyond basic RAG requires thoughtful decisions about ingestion, indexing, vector search, keyword retrieval, metadata filtering, data freshness, authorization, ranking, and provenance. This session explores how to design the data layer for production agentic applications on Azure. We will examine structured and unstructured information, hybrid retrieval, Azure AI Search, enterprise data sources, evidence ranking, citations, and secure data access. Attendees will learn how to build retrieval architectures that provide agents with useful context while maintaining traceability, security, freshness, and operational control.
Focus: Building the data and retrieval foundation behind trustworthy enterprise AI agents: ingestion, hybrid search, vector retrieval, metadata, freshness, authorization, grounding, and citations.
MCP Meets Kubernetes: Giving AI Agents Access to Cloud-Native Tools
AI agents become much more useful when they can interact with real systems instead of relying only on what an LLM already knows. Model Context Protocol (MCP) provides a standardized way to connect agents with external tools, data, and services.
In this practical session, we’ll combine MCP, Google’s open-source Agent Development Kit (ADK), and Kubernetes to build an AI agent capable of investigating a cloud-native environment. Rather than simply asking an LLM, “Why is my application failing?”, we’ll allow the agent to use controlled tools to inspect Kubernetes resources, gather relevant context, and use Gemini to reason about what it discovers.
We’ll walk through the MCP client/server model, connecting MCP tools to an ADK agent, containerizing the components, and safely exposing Kubernetes capabilities to the agent. We’ll also discuss permissions and why agents should receive only the access they actually need.
**What You’ll Learn:**
* Understand MCP architecture
* Connect MCP tools to Google ADK
* Give agents controlled access to Kubernetes
* Inspect workloads and resources using an agent
* Containerize MCP and agent components
* Apply safe tool-access and permission patterns
From Search to Prevention: Evidence-Grounded AI Agents for Personalized Health Navigation
Instead of asking, “Which doctor should I see?”, the system helps a person navigate preventive health information over time. The agent retrieves trustworthy evidence, considers the user's stated goals/context, identifies relevant preventive-care information, explains the evidence, and directs the user toward appropriate professional care rather than attempting diagnosis.
Preventive healthcare increasingly begins outside the clinic, where individuals search for information about screenings, wellness, risk factors, and when to seek professional care. Generative AI can personalize these interactions, but unsupported recommendations and unclear evidence provenance create significant trust challenges.
This session presents an evidence-grounded agentic AI approach for personalized preventive health navigation. Specialized agents decompose health and wellness questions, retrieve authoritative biomedical evidence, rank sources by relevance and authority, and generate transparent responses linked to their supporting evidence. The framework emphasizes prevention, education, appropriate escalation to professional care, and preservation of human decision-making rather than autonomous diagnosis.
Using practical preventive-health scenarios, the session demonstrates how agentic AI can move beyond one-time generative answers toward longitudinal, evidence-aware health navigation while maintaining provenance, uncertainty, and safety boundaries.
From Clinical Question to Evidence: Multi-Agent AI for Transparent Healthcare Evidence Synthesis
Clinicians and care teams must navigate growing volumes of biomedical literature, provider information, and other healthcare data while making time-sensitive decisions. Generative AI can accelerate information synthesis, but conventional approaches can obscure where information came from and whether citations actually support the response.
This session presents a multi-agent AI architecture that separates query planning, healthcare retrieval, evidence ranking, and grounded response generation. Through a working healthcare AI prototype, the session demonstrates how specialized agents can retrieve provider and biomedical evidence, preserve provenance, rank sources, and synthesize information while maintaining human oversight. The approach explores how agentic AI can assist healthcare professionals without turning an LLM into an autonomous clinical decision-maker.
From Prompt to Production: Deploying Gemini-Powered Apps on Kubernetes
Building an AI application locally is easy. Running it reliably in production is where cloud-native engineering begins.
In this hands-on session, we’ll take a simple Gemini-powered application from a local prompt to a containerized, production-ready workload running on Kubernetes. We’ll explore how Docker and Kubernetes provide the infrastructure needed to package, configure, deploy, scale, and operate modern AI applications without requiring a complex machine-learning platform.
Through a practical demo, we’ll containerize the application, create Kubernetes Deployments and Services, securely manage API keys using Secrets, configure the application with ConfigMaps, add health checks, and explore horizontal scaling. We’ll also discuss how the same open-source architecture can move from a local Kubernetes environment to Google Kubernetes Engine (GKE).
**What You’ll Learn:**
* Build a simple Gemini-powered application
* Containerize AI workloads with Docker
* Deploy and expose applications on Kubernetes
* Manage configuration and secrets securely
* Add health checks and scaling
* Understand the path from local Kubernetes to GKE
Can You Trust Your Healthcare AI Agent? Stress-Testing Grounding, Citations, and Tool Use
Healthcare AI is rapidly moving beyond chatbots toward retrieval-augmented and agentic systems that search external sources, call tools, rank evidence, and generate recommendations. But a response can appear well grounded while still relying on poor evidence, incorrect citations, or inappropriate tool behavior.
This session presents a practical framework for stress-testing agentic healthcare AI across evidence provenance, retrieval quality, citation fidelity, and tool use. Using a working healthcare discovery prototype, we demonstrate how controlled conflicting and low-quality evidence can expose weaknesses in retrieval and reasoning pipelines, and how safeguards such as authority-aware ranking, provenance tracking, citation verification, and evidence conflict detection can improve trustworthiness. The session focuses on what organizations should evaluate before moving healthcare AI systems from experimentation into real-world use.
Build Your Own AI Research Agent with Google ADK
Traditional chatbots wait for a prompt and generate an answer. AI agents can go further: they can search for information, use tools, reason across multiple sources, and execute multi-step research workflows.
In this session, we’ll build a practical AI research agent using Google’s open-source Agent Development Kit (ADK). Starting with a research question, our agent will determine what information it needs, invoke search and other tools, analyze the retrieved information with Gemini, and synthesize the findings into a useful, grounded response.
We’ll explore the core building blocks of agentic applications—agents, models, tools, sessions, orchestration, and grounding—without introducing unnecessary architectural complexity. The completed agent will then be containerized so the same project can run locally or as a cloud-native workload.
**What You’ll Learn:**
* Understand LLMs vs chatbots vs agents
* Build an agent with Google ADK
* Connect agents to search and external tools
* Create multi-step research workflows
* Ground AI responses with retrieved information
* Containerize an ADK application
Beyond Prompting: Teaching Healthcare Professionals to Evaluate and Trust AI Agents
Healthcare professionals increasingly interact with AI systems that retrieve evidence, call external tools, synthesize information, and recommend next steps. Effective AI education therefore requires more than teaching prompt engineering. Clinicians, researchers, and students must learn how to evaluate where an AI answer came from, whether its citations support its claims, how external tools were used, and when human verification is required.
This session presents a practical, case-based framework for building AI evidence literacy in the healthcare workforce. Using an evidence-grounded healthcare agent as a teaching environment, participants learn to inspect retrieval, provenance, evidence quality, citations, uncertainty, and agent behavior. The session proposes a competency model for preparing healthcare professionals to use agentic AI critically, safely, and effectively.
Beyond Generative Search: Building Evidence-Grounded Agentic AI for Healthcare Discovery
As patients increasingly use generative AI to discover providers, services, and health information, healthcare search is shifting from ranked links toward AI-generated recommendations. But healthcare discovery requires more than fluent answers—it requires trustworthy evidence, source transparency, and traceability.
This session presents an evidence-grounded agentic AI framework that decomposes patient questions into targeted searches, retrieves healthcare provider data and biomedical literature from authoritative sources, ranks evidence, and generates responses grounded in verifiable citations. Using a working healthcare discovery prototype, the session demonstrates how specialized AI agents can separate planning, retrieval, evidence ranking, and response generation. The goal is to explore how agentic AI can support a more transparent and trustworthy future patient journey while reducing the risks of hallucination and unsupported recommendations.
Beyond Chatbots: Running AI Agents on Kubernetes
AI is moving beyond chatbots toward agents that can reason, use tools, call APIs, and execute multi-step tasks. In this session, we’ll build an AI agent using Google’s open-source Agent Development Kit (ADK), containerize it with Docker, and deploy it on Kubernetes.
Through a practical open-source demo, we’ll explore the journey from a locally running agent to a scalable, cloud-native agentic application.
What You’ll Learn:
How AI agents differ from traditional chatbots and LLM applications
How to build an agent using Google’s open-source ADK
How to containerize AI agents with Docker
How to deploy and manage agents on Kubernetes
How agents connect to LLMs, tools, APIs, and other agents
How to use Kubernetes Services, ConfigMaps, Secrets, and health checks for agent workloads
How to scale and observe AI agents in production
How to get started with the open-source demo and extend it for your own use cases
Through a practical open-source demo, we’ll follow an agent from local development to a Kubernetes deployment and explore how agents connect with LLMs, tools, APIs, and other agents.
Attendees will leave with a practical architecture and an open-source example they can use to start running AI agents on Kubernetes.
Top 15 Security Mistakes Startups Make on AWS (and How to Fix Them)
Startups move fast—and that's exactly why security is often overlooked until it's too late. From accidentally exposing Amazon S3 buckets to granting overly permissive IAM roles, small configuration mistakes can lead to costly breaches, compliance issues, and downtime.
In this session, we'll explore 15 of the most common AWS security mistakes startups make while building and scaling their applications on the AWS tech stack. You'll learn why these issues happen, understand the real-world risks they introduce, and discover practical, AWS-native solutions to prevent them.
Whether you're a founder, developer, DevOps engineer, Solutions Architect, or cloud enthusiast, this talk will provide actionable best practices that you can apply immediately to build a more secure AWS environment.
What you'll learn:
1. The 15 most common AWS security mistakes startups make
2. How attackers exploit common cloud misconfigurations
3. AWS security best practices for IAM, networking, storage, secrets, and encryption
4. How to use services like AWS IAM, AWS Organizations, AWS CloudTrail, Amazon GuardDuty, AWS Config, AWS Security Hub, AWS WAF, AWS KMS, and AWS Secrets Manager to improve your security posture
5. Practical techniques to secure applications without slowing down development
6. A simple security checklist every startup should implement before launching to production
By the end of this session, you'll leave with practical guidance, a security-first mindset, and a roadmap to help protect your AWS workloads as your startup grows.
Modern AI-powered Healthcare Data Lake on AWS
Healthcare organizations generate vast amounts of data from electronic health records (EHRs), medical imaging, laboratory systems, wearable devices, and patient applications. Turning this data into actionable insights requires a modern, secure, and scalable data platform.
In this session, we'll explore how to build an AI-powered healthcare data lake on AWS that centralizes structured and unstructured data while enabling advanced analytics and generative AI use cases. You'll learn how AWS services can be combined to securely ingest, catalog, govern, analyze, and visualize healthcare data, while maintaining compliance and preparing it for AI-driven applications.
Whether you're a cloud architect, data engineer, developer, healthcare technologist, or AI enthusiast, this session will provide practical architecture patterns and best practices for designing modern healthcare data platforms on AWS.
What you'll learn:
1. How to design a scalable healthcare data lake architecture on AWS
2. Best practices for ingesting, storing, and governing healthcare data using Amazon S3, AWS Glue, AWS Lake Formation, and Amazon Athena
3. Approaches to building secure, compliant data platforms with encryption, IAM, auditing, and fine-grained access controls
4. How to integrate generative AI and analytics using Amazon Bedrock, Amazon QuickSight, and other AWS AI services
5. Real-world healthcare use cases such as patient analytics, clinical dashboards, document summarization, and population health insights
6. Cost optimization, monitoring, and operational best practices for production-ready healthcare data platforms
By the end of this session, you'll understand how to build a secure, scalable, and AI-ready healthcare data lake on AWS that transforms raw healthcare data into meaningful insights and intelligent applications.
From RAG to Production: How SMBs Can Grow Securely with GenAI on AWS
Generative AI is transforming how small and medium-sized businesses (SMBs) operate, from intelligent customer support and document search to knowledge management and business automation. While building a proof of concept is relatively straightforward, deploying a secure, scalable, and production-ready AI application presents a different set of challenges.
In this session, we'll explore how to take a Retrieval-Augmented Generation (RAG) application from prototype to production using AWS. We'll walk through the key architectural components, discuss security and governance best practices, and examine how AWS services can be used to build reliable, cost-effective GenAI solutions that scale with business needs.
Whether you're a developer, solutions architect, startup founder, or cloud engineer, you'll gain practical insights into designing and deploying production-ready GenAI applications on AWS.
What you'll learn:
1. What Retrieval-Augmented Generation (RAG) is and why it's a preferred approach for enterprise and SMB AI applications
2. How to design a secure, scalable RAG architecture on AWS
3. Best practices for document ingestion, embeddings, vector search, and prompt orchestration
4. How to use services such as Amazon Bedrock, Amazon S3, AWS Lambda, Amazon API Gateway, Amazon Cognito, and Amazon OpenSearch Service to build end-to-end GenAI applications
5. Security considerations including IAM, encryption, access controls, guardrails, and protecting sensitive business data
6. Strategies for monitoring, optimizing costs, and operating GenAI workloads in production
7. Common pitfalls when moving from proof of concept to production and how to avoid them
By the end of this session, you'll understand the architectural patterns, security principles, and operational best practices needed to build production-ready GenAI applications on AWS that deliver business value while maintaining security, reliability, and scalability.
Cloud Careers: Combining AWS, Security, and AI Skills - Founder's Journey
The cloud industry is evolving faster than ever, and the most valuable professionals are no longer experts in just one domain. Organizations are looking for people who can combine cloud architecture, cybersecurity, and AI to build secure, scalable, and intelligent solutions.
In this session, I'll share my personal journey—from learning AWS and cloud technologies to building expertise in security, exploring generative AI, and eventually starting my own company. Along the way, I'll discuss the lessons learned, the challenges faced, and the skills that have had the greatest impact on my career.
Whether you're a student, aspiring cloud engineer, developer, security professional, or someone looking to transition into cloud and AI, this session will provide practical advice on building the right skills, gaining hands-on experience, and creating a long-term career roadmap.
What you'll learn:
1. Why AWS, cybersecurity, and AI are a powerful combination for today's technology careers
2. A practical learning roadmap for building cloud, security, and AI skills
3. How to gain real-world experience through projects, certifications, and community involvement
4. The importance of security-first thinking when designing cloud solutions
5. Lessons learned from transitioning from engineer to founder
6. Common career mistakes to avoid and strategies for continuous growth
7. How to build a strong professional brand through open-source contributions, technical content, public speaking, and networking
By the end of this session, you'll have a clear roadmap for developing in-demand technical skills, growing your career in cloud and AI, and understanding how continuous learning and community involvement can open doors to new opportunities, including entrepreneurship.
The Hidden Layer in AI Healthcare: Who Controls the LLM Recommendation?
Patients are increasingly turning to LLMs and AI systems for medical guidance before traditional healthcare channels, but these systems do not simply generate answers; they actively shape recommendations by ranking, filtering, and prioritizing medical information, providers, and treatments. This introduces a hidden recommendation layer inside AI systems where outcomes are influenced by training data, retrieval pipelines, system prompts, and safety alignment layers. In healthcare, this creates a critical security and governance challenge, as these recommendation pathways are not transparent or auditable. The talk explores how LLM recommendation stacks work, where control points exist, and why emerging risks such as prompt injection, content poisoning, and recommendation manipulation directly impact patient safety. It also highlights the governance gap in AI-driven healthcare decisions, raising the core question of who ultimately controls what AI recommends and how that control is exercised.
Hidden Facts: Training Data, RAG, Prompts, and Safety Layers Behind AI-Driven Healthcare Advice
Patients are increasingly turning to AI systems and large language models (LLMs) for medical guidance before engaging with traditional healthcare channels. However, these systems do not merely answer questions—they shape recommendations by selecting, ranking, filtering, and prioritizing information about symptoms, treatments, medications, and healthcare providers.
Behind every AI-generated response lies a set of hidden control points: training data, retrieval-augmented generation (RAG) pipelines, system prompts, and safety and alignment layers. These components collectively determine what information is surfaced, what is omitted, and how recommendations are framed. Despite their growing influence on healthcare decisions, these mechanisms remain largely opaque and difficult to audit.
This session explores how these hidden recommendation layers operate, where control and influence exist within AI systems, and why emerging risks such as prompt injection, content poisoning, and recommendation manipulation have direct implications for patient safety and trust. It also examines the widening governance gap created by AI-driven healthcare advice, where transparency, accountability, and auditability have not kept pace with adoption.
The discussion ultimately addresses a critical question for healthcare organizations, regulators, and technology providers: Who controls what AI recommends, and how can these hidden control points be made transparent, auditable, and accountable in healthcare?
Healthcare governance gap: lack of auditability and transparency in AI-driven advice
As patients increasingly turn to large language models (LLMs) and AI assistants for medical guidance before consulting traditional healthcare channels, a new governance challenge is emerging. These systems do far more than generate answers—they actively influence decisions by ranking, filtering, and prioritizing medical information, treatment options, and healthcare providers.
Beneath every AI-generated response lies a complex recommendation stack shaped by training data, retrieval mechanisms, system prompts, safety guardrails, and alignment layers. This hidden recommendation layer determines what information is surfaced, suppressed, or emphasized, yet it remains largely opaque and difficult to audit.
The session examines how AI recommendation systems in healthcare operate, where control points exist, and why vulnerabilities such as prompt injection, content poisoning, and recommendation manipulation represent more than technical risks—they directly impact patient safety, trust, and clinical outcomes. It also explores the growing governance gap created by AI-driven healthcare advice, where accountability and transparency mechanisms have not kept pace with adoption.
Ultimately, the discussion raises a fundamental question for healthcare leaders, regulators, and technologists: Who controls what AI recommends in healthcare, and how can those decisions be made transparent, auditable, and accountable?
Your AI Agent Works in a Demo. Now Put It in Production: 7 Engineering Problems Nobody Warns You
Building an AI agent that works in a demo is easy. Making it "reliable, secure, observable, and cost-effective in production" is where the real engineering begins.
In this practical, architecture-focused session, we’ll explore "7 challenges that emerge when AI agents move from prototype to production":
1. Identity & Authorization — Who is the agent acting as?
2. Secrets & Credentials — How should sensitive access be managed?
3. State & Memory — What happens when workflows fail halfway?
4. Tool Reliability — Handling retries, timeouts, and duplicate actions.
5. Observability — Tracing decisions across models, tools, and APIs.
6. Cost & Latency — Controlling model calls and response times.
7. Security & Guardrails — Limiting unsafe actions and blast radius.
We’ll connect these challenges to practical engineering patterns and a production-ready architecture.
Tulsa AWS Community Day Sessionize Event
Please note that Sessionize is not responsible for the accuracy or validity of the data provided by speakers. If you suspect this profile to be fake or spam, please let us know.
Jump to top