Aman Sharma
Cofounder Lamatic.ai, Building Florida AI Community @AI Collective
Miami, Florida, United States
Actions
Aman is a serial AI entrepreneur with 3 successful exits and over a dozen digital/AI products. His journey began as an ML researcher at TU Wien, where he pioneered CNN for large-scale recommendation systems. A pillar of the South Florida tech scene, Aman leads the AI Collective Miami Chapter. Aman has delivered 100+ tech talks with his signature Miro Diagram style storytelling, which has inspired thousands of founders, executives, and developers worldwide, establishing him as a two-time Most Active Speaker and respected AI innovator.
Badges
Area of Expertise
Topics
Multi-Agent Deadlock & Consensus: When Agents Disagree (And What to Do About It)
Shipping a single agent is hard. Shipping multi-agent systems that coordinate reliably is exponentially harder. When Agent A's output becomes Agent B's input becomes Agent C's decision, what happens when they disagree? How do you detect deadlock? How do you resolve conflict without cascading failures? Most agent frameworks focus on chaining: A→B→C linear pipelines. Enterprise workflows demand orchestration: multiple agents deciding simultaneously, negotiating outcomes, recovering from failures. This is where distributed systems meets agent behavior. Recent production deployments reveal new failure modes: consensus loops (agents pass back and forth endlessly), deadlock cascades (A waits for B, B waits for C, C waits for A), inconsistent views (agents decide on stale info), timeout ambiguity (fail or still thinking?). This talk explores: The multi-agent problem space (consensus, deadlock, consistency, Byzantine agents). Detection patterns (timeouts, circular dependency analysis, state divergence). Resolution strategies (voting & arbitration, process-level agreements, timeout hierarchies, rollback & replay, Byzantine tolerance). Production case studies (E-commerce order processing, autonomous research coordination, multi-agent audits). This draws from distributed systems theory but applies it directly to agentic workflows.
Most agent frameworks chain A→B→C. Enterprise workflows demand orchestration: multiple agents deciding simultaneously, negotiating outcomes, recovering from failures. New failure modes: consensus loops, deadlock cascades, inconsistent views. This talk explores detection patterns and six resolution strategies (voting, arbitration, formal consensus, timeouts, Byzantine tolerance, rollback) from distributed systems applied to agents. Case studies: E-commerce orchestration, autonomous research, multi-agent audits.
Shift to Software 3.0 : Turning SaaS Agentic
We’re entering the era of Software 3.0, where SaaS becomes agentic—autonomous, adaptive, and capable of acting on user intent. This session charts the shift from Software 2.0 to 3.0, exploring new UX patterns, practical playbook principles, and technical strategies for building agentic SaaS. Through interactive storytelling and real-world case studies, attendees will gain clear, actionable insights on how to harness this shift to accelerate development and deliver next-generation SaaS experiences.
Stateful Agents at Scale: Memory, Consistency, and the Illusion of Continuity
Long-running agents are broken. They forget critical facts mid-task, lose reasoning coherence across sessions, and degrade as interactions accumulate. The problem: we treat agents like stateless APIs, not stateful systems. Context windows are expanding (1M tokens), yet agents still struggle with context collapse, information loss, and irreversible decisions. A customer support agent processes 50 interactions but can't remember the customer's stated preference from conversation #3. A research agent reads 200 papers but loses track of critical early citations when new evidence emerges. The frontier is designing memory systems that span fixed-size context + external persistence + intelligent retrieval. Recent breakthroughs (ReMemR1, Anthropic's context engineering, MemGPT patterns) show how agents can maintain coherence over hundreds of thousands of tokens. This talk explores: The memory problem (context rot, hallucination, forward-only processing, irreversibility). Emerging architectures (ReMemR1: +6-10% accuracy, MemGPT core blocks, context engineering, sleep-time compute). Production challenges (consistency models, eviction, retrieval semantics). Case studies (multi-session agents that improve). This talk bridges academic research to production practice, giving engineers the frameworks to build agents that actually learn and remember.
Long-running agents forget critical facts, lose reasoning coherence, and degrade as interactions accumulate. Context windows expand (1M tokens) yet agents struggle with context collapse. Recent breakthroughs (ReMemR1, Anthropic context engineering, MemGPT patterns) show how agents maintain coherence over hundreds of thousands of tokens. This talk explores the memory problem, four emerging architectures, production challenges, and case studies of agents that improve with every interaction.
The Hidden Cost of Model Diversity: Managing 20+ LLM APIs in Production
Every enterprise shipping production agents eventually hits a wall: they need multiple models. GPT-4 for reasoning, Claude for compliance, Llama for cost, local models for privacy, Gemini for multimodal. The problem: managing 20+ LLM APIs isn't engineering—it's an operational crisis. Most teams start with one model. Cost pressures force switching. Compliance mandates specific vendors. Latency requires fallbacks. Suddenly you're managing: 20+ API contracts (versioning, rate limits, auth), cost tracking per model per task, fallback chains, compatibility matrices, infrastructure sprawl. Result: engineering spends more time on model operations than agent logic. This talk explores: Operational cost of diversity (hidden burden, engineering tax, complexity). API contract management (versioning, normalization, testing). Cost attribution & optimization (multi-tenant tracking, selection, waste). Reliability patterns (fallback hierarchies, circuit breakers, queuing, degradation). Vendor management (lock-in mitigation, negotiation, SLAs). Production case studies: Finance (12 models, 35% engineering time), Healthcare (8 models, compliance nightmare solved), Enterprise ($2M/year savings via routing). We show how to abstract diversity so engineers focus on agent logic, not infrastructure.
Every enterprise needs multiple models: GPT-4 for reasoning, Claude for compliance, Llama for cost, local for privacy. Managing 20+ APIs is an operational crisis: API contracts, cost tracking, fallback chains, compatibility matrices, infrastructure sprawl. This talk explores the operational tax of diversity, API abstraction, cost attribution, reliability patterns, and case studies where model routers save $2M/year. How to abstract diversity behind clean APIs so engineers focus on agent logic.
The Lean GenOps playbook: Building GenAI app faster and scalable
This comprehensive master class is designed to guide anyone (non-AI and noncoders) from the initial ideation phase of GenAI features through to the successful deployment of scalable applications in production. Here’s a detailed, pointwise breakdown of what the playbook covers:
1. Introduction to GenAI and LLMs: Quick dive into what GenAI and LLMs are, emphasizing their relevance today.
From Idea to Concept: Rapid ideation techniques for GenAI features.
2. Choosing the Right Model: Key points on selecting LLMs for your project’s needs.
3. Integrating your data: Selecting data and Overview of how vector databases support LLM applications.
4. LLM Training Essentials: Brief on training and fine-tuning LLMs for custom needs.
5. Simplifying LLMOps:Introduction to managing LLM applications with a focus on practical tools.
6. Building Your Application:Key steps to integrate LLMs into your app’s front and back end.
7. Deployment and Testing:Strategies for deploying and testing your application efficiently.
8. Monitoring, Maintenance, and Scaling:Tips for keeping your application running smoothly and scaling up.
9. Wrap-Up: Best practices and ongoing improvements
Sleeper Agents: Asynchronous Reasoning That Decouples Latency From Accuracy
The test-time compute revolution promised better agent decisions—at the cost of latency. But what if we could have both? Sleeper agents represent a paradigm shift: autonomous systems that reason during idle time (non-inference periods), precomputing insights that enable instant responses while maintaining or improving quality. Unlike test-time scaling, which forces users to wait, sleeper agents push reasoning to background compute. A lightweight serve agent answers queries instantly using precomputed context. Result: 1/5 token reduction with 15% accuracy improvements, and 2-3x cost amortization on multi-turn conversations. But this raises new challenges: How do you precompute reasoning without knowing future queries? What representations persist best in memory? How do you prevent hallucination propagation? This talk explores architecture (memory abstraction, precomputation, coordination), challenges (speculative reasoning, eviction, staleness, hallucination), implementation patterns (sleep triggers, refresh mechanics), and production case studies (latency reduced 40%, costs cut 50%). This is frontier infrastructure—barely covered in academic literature, no conference talks to date.
Test-time compute improved agent decisions—at the cost of latency. Sleeper agents invert this: autonomous systems reasoning during idle time, precomputing insights for instant responses. Result: 1/5 token reduction, 15% accuracy gains, 2-3x cost amortization. This talk explores three architectural challenges (speculative reasoning, memory representation, context refresh), production patterns, and case studies of teams reducing latency 40% while cutting costs 50%.
Agent as Code - Complete Improvement
Agents are software—but most teams build them as notebooks: tightly coupled to specific LLMs, brittle infrastructure dependencies, impossible to version, unreproducible, and nightmarish for collaboration.
Traditional agent development entangles logic with API calls, specific model names, hardcoded credentials, database assumptions. Change the LLM provider? Rewrite half the code. Move from cloud to edge? Redesign everything. Add a second engineer? Merge conflicts and inconsistency.
AI Agent Workshop
Orlando Code Camp 2025 Sessionize Event
eMerge Americas
Finalist Pitch
2025 Florida Venture Capital Conference
Finalist Presentation
React Global Online Summit'25 AI Edition
Beyond Chatbots
AI Builders Night
TechCrunch Disrupt Startup Battlefield
Agile + DevOpsDays Des Moines 2024 Sessionize Event
DevOps Days Tampa Bay 2024 Sessionize Event
GenAI Meetup Delhi
AI Uncorked
Building GenAI apps : Zero to Hero GenAI playbook
CTO Insights
GenAI 0 to hero: Playbook to Building GenAI Apps with LLM
AI DevSummit 2024 Sessionize Event
DevOpsDays Philadelphia 2024 Sessionize Event
SW2CON
Beyond Chatbots: Building Embedded SAAS experiences with GenAI
CNCF ML Bootcamp
LLM Fundamentals 0 to 1
Atlanta Cloud Conference 2024 Sessionize Event
Michigan Technology Conference 2024 Sessionize Event
DDD Melbourne 2024 Sessionize Event
Data.TLV Summit 2024 Sessionize Event
Orlando Code Camp 2024 Sessionize Event
THAT conference Texas
The Lean DevOps Playbook
Building a GenAI product
A session on how to build a GenAI product
National Startup Day at NSIT
This Even was about sharing my startup Journey in the session named Zero to Hero and also Judging a Startup pitching competition.
90DaysOfDevOps - 2024 Community Edition Sessionize Event
Google DevFest Ludhiana 2023 (In-Person) Sessionize Event
GDG Ahmedabad DevFest 2023 Sessionize Event
Symposium on Computational Gastronomy
Food X AI : Leveraging LLM for food API
AI DevWorld 2023 Sessionize Event
GDG Athens DevFest 2023 Sessionize Event
ChefConf.23 Germany Sessionize Event
DevOps Days Tampa Bay 2023 Sessionize Event
WorldFestival 2023 Sessionize Event
SEI Secure Software by Design Sessionize Event
Global Mobile summit
Represented and discuss community landscape from AMP side
India web community lead summit
Represented and discuss community landscape from AMP side
Cyber Security Global Summit 2021 DAY 1 - Entry-level track
DevSecOps Days Pittsburgh 2023 Sessionize Event
Modern Frontends Live! 2022 Sessionize Event
NFT.London 2022 by NFT.NYC Sessionize Event
Localhost Conference 2022 Sessionize Event
Global AI Developers Days Sessionize Event
Devopsdays Eindhoven 2022 Sessionize Event
Mêlée Tech - 2022 Sessionize Event
Web Developer Conference '22 Sessionize Event
Web 3 Conference
(Democratized, Decentralized research ecosystem with NFT and web3)
Sharing the idea of using web3 for a world of decentralized research at web3 conf
Twitter Space event(Web Application Firewall 101 )
Discussing about web application firewalls and how to get started with them with NGINX engineering leed
Accessibility Days 2021 Sessionize Event
Cyber Security Global Summit 2021
(Security and UX hand in hand | Link)
How to provide better security along with good UX at Cybersecurity global summit
Accessibility Days(Accessiblity with AMP)
How can we build more accessible internet using AMP framework
React JS summit
diminishing the line between a web and native apps using ux and pwa practices at React global summit
Dnn Summit
making optimized web project with amp and using it with DNN CMS
DNN Virtual Summit 2021 Sessionize Event
Global IT Security Conference
using UX to your advantage to create secure apps at Gloabl IT security summit
Hour of Code | ACM Student Chapter IIT(ISM),Dhanbad
IIT ISM Dhanbad
Belpy
How to plan and use datascience project to your advantage for better ROI at Belpy 2021
Pycon Indonesia
A talk on managing projects and competing them on time at pycon indonesia
Pyjama Conference
A talk on managing projects and competing them on time at pyjamas conference
Pycon India
A talk on managing projects and competing them on time at pycon india
Python Global Summit
A talk on managing projects and competing them on time at python global summitA talk on managing projects and competing them on time at python global summit
Devfest Great NorthDevfest Great North
creating performant and optimized web project with amp at Devfest great north
IGDTUW
mythbusting cloud and explaining easier entrance to start cloud adoption and learning at IGDTUW
Jamstack Berlin
Using amp in JAMstack project at a meetup in jamstack berlin in collaboration with contentful
Doing a web startup the right way
a meetup session guiding entreprenuer about scaling there web startup at mobile-web.dev
GDG Cloud Ahemdabad
introduction to AMP and how it can help in making web application faster at GDG cloud ahemedabad meetup group
Web Dev Live
Vast possibilties and landscape of AMP at WEB Dev Live
Mobile-Web.dev
A talk on using AI on web application at community meetup of mobile-web.dev
Fossasia Summit
A talk on making better performant web apps using the concept of AMP and PWA at Fossasia sumit singapore
OpenTech Summit
A talk on making better performant web apps using the concept of AMP and PWA at opentech summit Delhi
OpenTech Summit
A talk on making better performant web apps using the concept of AMP and PWA at opentech summit Srilanka
Wordfest
Using AMP to Developers advantage to create performant and fast websites. Also introduced Wordpress way of doing AMP.
About mobile-web.dev and AMP
Introduction to AMP and our mobile-web.dev community for awareness
Online Meetup
A talk on making better performant web apps using the concept of AMP and PWA
Online Meetup
Kickstarted AMP community interaction in India which was a seed for mobile-web.dev community
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