AWS Community Day Bay Area 2026
Selected talks
Recent recordings
AI Performance Engineering Meetup 2026
From Agent Traces to Agent Training: Closing the Continuous Learning Loop
Watch the recordingAI Engineer World's Fair 2026
From Agent Traces to Agent Simulations: The Next Era of Agent Evaluation
Watch the talkAWS Community Day Security Edition 2026
Isolation vs Autonomy: Choosing Sandboxes for Terminal AI Agents
Watch the talkSelected work
AgentOps Simulation Loop
A reference implementation connecting synthetic billing-agent traces to reproducible scenarios, isolated Harbor trials, and state verification.
View repositoryOffline evals for coding agents
An educational demo of coding-agent evaluation with fixed Python tasks, reproducible checks, and saved conversations, patches, and reports.
View repositoryStepFunctions2Processing
AWS Step Functions processing examples using Lambda, Batch, and Fargate.
View repositoryIndustry work & case studies
External case studies and customer perspectives from my work with Snorkel AI’s platform team on agent simulation and evaluation.
Realistic agent simulations with Modal
Modal explores how Snorkel AI uses CPU and GPU sandboxes to recreate complex evaluation environments, with my perspective on making agent simulations reflect real workloads.
Read Modal case studyScaling agent evaluation with Daytona
Daytona profiles my team’s work at Snorkel AI to provision configurable, isolated sandboxes and run agent evaluations in parallel.
Read Daytona case studyFeatured in AWS launches
-
My customer quote discusses how isolated execution and usage-based pricing could support Snorkel AI’s agent simulations and evaluation workloads.
-
My customer quote describes how Snorkel AI’s teams, agents, and evaluators can use file-system access to S3 without duplicate data copies or synchronization waits.
More projects
Lambda packages
Packaged environments for AWS Lambda that enable use of various Python libraries.
View repository
Amazon skills
Several dynamic Alexa skills that involve scraping data from online sources.
Explore the projectBooks & courses
Authored book
Hands-On Serverless Deep Learning with TensorFlow and AWS Lambda
A practical guide to deploying TensorFlow models with AWS Lambda, building APIs, and assembling serverless processing pipelines.
Contributed chapter
97 Things Every Cloud Engineer Should Know
Contributor: Chapter 23 - Cloud Processing Is Not About Speed
Perspectives on cloud engineering from practitioners across the industry, including my chapter on cloud processing.
More books & courses
Contributed chapter
97 Things Every Data Engineer Should Know
Contributor: Chapter 67 - The Data Pipeline Is Not About Speed
Practical lessons in data engineering, including my chapter on the purpose of data pipelines.
Course
Serverless Deep Learning with TensorFlow and AWS Lambda
A course on deploying deep learning models with AWS serverless infrastructure.
Writing
Coding Agents Escaped the IDE. Bring Them to Your App.
Serverless compute for LLM — with a step-by-step guide for hosting Mistral 7B on AWS Lambda
Guide for Running Llama 2 Using LLAMA.CPP on AWS Fargate
Making LLMs Scalable: Cloud Inference with AWS Fargate and Copilot
How to serve deep learning models using TensorFlow 2.0 with Cloud Functions
More writing
Top 20 AWS re:Invent 2024 breakout sessions to stay on top of everything related to generative AI
Re:Invent 2023 attendee guide: Generative AI
Scalable Cloud inference endpoint using ONNX and AWS Fargate
How AWS Lambda SnapStart drastically reduces cold starts for Serverless Machine Learning Inference
ML No/Low Code services and Use Cases at Re:Invent 2022
12 MLOps breakout sessions I'm looking forward to at Re:Invent 2022
7 things to know before using AWS Panorama
7 MLOps breakout sessions I'm looking forward to at Re:Invent 2021
Machine Learning Inference on AWS Graviton2 Lambda Functions
ML infrastructure @ AWS Summit 2021
MLOps sessions @ re:Invent 2020
Training models using Satellite imagery on Amazon Rekognition Custom Labels
Using custom docker image with SageMaker + AWS Step Functions
Amazon CodeGuru Profiler for monitoring cloud applications
Serverless Deep Learning: Challenges and Benefits
Using the Serverless framework to deploy hybrid serverless/cluster workflows
Using TensorFlow and the Serverless Framework for deep learning and image recognition
Serverless tensorflow on AWS Lambda
About
Rustem Feyzkhanov
Engineering Manager, AI Platform · Bay Area, CA
I’m an engineering manager focused on AI platforms. I lead engineering teams and work on the infrastructure and developer tools behind production AI, with a particular interest in agent evaluation and cloud systems.
I’m an AWS AI Hero, author, and speaker. My work spans agent evaluation, serverless computing, and machine learning infrastructure. Earlier, I built systems for manufacturing inspection and satellite imagery—experience that continues to shape how I think about AI in production.
I wrote Hands-On Serverless Deep Learning with TensorFlow and AWS Lambda and regularly share what I learn through talks, courses, and articles.
