Reducing MTTR and Alert Fatigue with PyTorch-Powered Anomaly Detection

Large B2C systems generate a deluge of telemetry facts, ranging from CPU and memory records to logs, bug codes, and nuanced platform-specific alerts throughout respective areas.
This consultation provides a hands-on approach to anomaly detection using Python to improve MTTR and decrease alert fatigue by way of delivering precise, honest alerts. Drawing on actual-world experience in growing a wise incident management bot, we explore the way to process diverse, multivariate statistics streams from pods, services, and platforms like iOS, Android, and the web. The model correlates metrics throughout infrastructure layers and platforms to surface a significant alerts, complete with context, diagnostics, and counselled remediation steps proactively.
Topics covered encompass:
Correlating contextual alerts to lessen noise and alert fatigue
Boosting confidence in alerts through meaningful insights
Automating diagnostics to force faster RCA
In case you're running big scale, this session will share sensible, adaptable strategies to embed deep gaining of knowledge into your observability stack and transition from reactive monitoring to proactive decision-making.

Moral Frontiers of AI marketers in Healthcare: Navigating Autonomy, Obligation, and Equity

As artificial intelligence evolves from static prediction equipment into agentic systems able to make self-sustaining choices, the ethical complexity in scientific environments intensifies. This consultation delves into the emerging moral terrain where AI dealers—functioning as virtual triage aides, diagnostic companions, or digital care managers—at once affect vital patient outcomes.
Key dialogue factors consist of:
The ethical shift from AI as a passive recommender to an active player in decision-making, and the results for responsibility and legal responsibility.
Case research (each actual and hypothetical) illustrating the function of AI marketers in diagnostics, individualised treatment, and far-off health tracking.
The critical significance of robust oversight mechanisms, obvious decision common sense, and inclusive design to mitigate bias and ensure trustworthy consequences.
The emergence of recent interdisciplinary roles includes AI ethics liaisons, clinical set of rule inspectors, and transparency engineers.
This speaks advocates for a practical integration approach that preserves middle human values—like dignity, autonomy, and agreement—while leveraging the efficiencies of AI.
Introducing the CARE-GRID Framework: A Governance Blueprint for AI-driven Care retailers
CARE-GRID stands for:
Contextual Oversight • Auditability • duty Anchoring • fairness through layout • Guardrails • real-Time monitoring • Interpretability • decision Escalation
Framework additives:
Contextual Oversight:
sellers function below care-context parameters (e.g., pediatric vs. oncology care) and jurisdictional compliance protocols (e.g., HIPAA, GDPR).
Auditability:
Each choice is logged with transparent metadata—reasoning, version model, and information source traceability—to permit thorough submit hoc assessment.
Obligation Anchoring:
AI actions are tethered to responsible human stakeholders. For instance, marketers may additionally stumble on anomalies; however, escalation should be human-established.
Fairness using design:
fashions are educated and tested with demographic consultant records. Every day, disparity tests are used to identify and cope with bias.
Guardrails:
Predefined constraints save your agents from executing or suggesting risky or unauthorised movements (e.g., medicinal drug prescriptions outdoor the scope).
Actual-Time tracking:
Sellers are constantly evaluated for model float or emergent bias, the usage of live monitoring tools, and automated alert systems.
Interpretability:
AI sellers need to provide clean, layperson-friendly causes to aid shared information and informed decision-making.
Decision Escalation Protocol:
Retailers adhere to an established hierarchy—autonomous for low-threat actions, supervisory review of mild instances, and ethics committee referral for high-impact selections.
Case examine: Algorithmic Bias in population health control
UnitedHealth’s Optum set of rules (2019)
trouble: The system disproportionately prioritised white patients for additional care, primarily based on ancient spending, marginalising patients with greater unmet needs.
Underlying Flaw: Reliance on healthcare spending as a proxy for health need embeds systemic disparities.
Corrective actions:
Shifted to clinical signs, together with comorbidity profiles and lab results to the degree of real care requirements.
Included equity-aware modelling and demographic impact scoring.
Instituted an ordinary bias audit workflow for continued oversight and adjustment.
This session gives an ahead-searching but grounded attitude on responsibly embedding AI retailers into healthcare transport, ensuring that innovation aligns with ethics, fairness, and human-centred care.

Incident Intelligence powered by AI Agents: Reducing MTTR with a Self-Healing Engineering Assistant

As platform scale and system complexities grow, so is the operational burden during outages. This session introduces how an AI Agent can be purpose-built for incident intelligence and self-healing system enablement — supporting Production Engineering, On-Call Dev Engineers, Engineering leaders and Exec Engineering Leadership by automatically generating contextual summaries, analyzing root causes, and offering prescriptive next steps.

Beyond just alerts, this AI Agent serves as a real-time co-pilot that pulls signals from deployments, telemetry, infrastructure, and runtime JVM artifacts like JFR, Heap Dumps, and Thread Dumps, config management, experiments launched and then distills and correlates them to determine the scope and probable cause of the issue thus keeping the MTTR and MTTD low

Data-Driven & Practitioner-Centric: An Engineering Leader’s Playbook for GenAI Adoption Metrics

In the rush to adopt generative AI tools like GitHub Copilot, many engineering leaders face a challenge: measurement. This talk presents a data-driven, role-sensitive metrics framework that enables organizations to move from experimentation /adoption to structured, scalable GenAI transformation.
This framework is categorized into:
• Usage & Engagement: Are Teams using GenAI tools? Where does usage drop, and when?
• Productivity & Throughput: How much time is saved? How are PR metrics evolving or degrading?
• Code Quality: Is GenAI-generated code as maintainable and bug-free as the code we write ourselves, and how do we know the auto-generated code is not introducing bugs?
• Role-Based Impact: Are junior engineers learning? Are seniors getting to focus on architecture?
Attendees will go out with:
1) Understanding of co-pilot telemetry
2) Pitfalls of adoptions and guardrails
3) Generating baseline metrics
This session is for Engineering Leaders, DevEx Teams to help them lead GenAI Initiatives with measurable impact

Continuous Improvement of Error Budgets to Business Wins: Driving Walmart Checkout Success with AI &

https://www.meetup.com/leaddev-meetup-sanfrancisco/events/307373895/?eventOrigin=group_upcoming_events

Machine Learning Summit 2025: Applied ML Engineering to GenAI and LLMs

https://www.eventbrite.com/e/machine-learning-summit-2025-applied-ml-engineering-to-genai-and-llms-tickets-1332848338259?aff=ebdsoporgprofile

Panel: Unifying teams for successful digital transformation: Strategies for engagement and collabora

https://www.digitaltransformation-week.com/northamerica/agenda/digital-transformation-in-action/

https://www.digitaltransformation-week.com/northamerica/speaker/kapil-poreddy/

– In this dynamic session, we will explore effective strategies for rallying teams around digital transformation initiatives. Participants will learn how to foster a culture of collaboration and engagement that drives successful outcomes.

– We will discuss the importance of clear communication, inclusive leadership, and shared vision in aligning diverse teams toward common goals. Attendees will also discover practical tools and techniques to overcome resistance to change, encourage cross-functional cooperation, and empower team members at all levels.