Uploaded May 2025 | Updated November 2026, 1 hour ago
[2025 - Day 3 - Lightning Talks] David Karam shares insights from building effective ranking systems with multi-dimensional functions that encode corpus topology and query streams, exploring how to merge signals like relevance, trustworthiness, and personalization at runtime. This talk offers valuable perspectives on recognizing emerging AI challenges as search problems in disguise, particularly relevant for engineers working on prompt optimization, ensemble selection, and reward modeling architectures inspired by decades of Google
Search work.
ABOUT THE SPEAKER:
David Karam, Co-Founder & CEO, Pi Labs -
ποΈ GET YOUR TICKET TO AI COUNCIL 2026 ποΈ
Meet the world's top AI infrastructure minds where architects of AI share what works. Three days of high-quality technical talks and meaningful interactions.
β aicouncil.com/sf-2026
β‘ FIND US:
X: https://x.com/AICouncilConf
LinkedIn: linkedin.com/company/aicouncilconf
Website: aicouncil.com
[2025 - Day 3 - Lightning Talks] David Karam shares insights from building effective ranking systems with multi-dimensional functions that encode corpus topology and query streams, exploring how to merge signals like relevance, trustworthiness, and personalization at runtime. This talk offers valuable perspectives on recognizing emerging AI challenges as search problems in disguise, particularly relevant for engineers working on prompt optimization, ensemble selection, and reward modeling architectures inspired by decades of Google
Search work.
ABOUT THE SPEAKER:
David Karam, Co-Founder & CEO, Pi Labs -
ποΈ GET YOUR TICKET TO AI COUNCIL 2026 ποΈ
Meet the world's top AI infrastructure minds where architects of AI share what works. Three days of high-quality technical talks and meaningful interactions.
β aicouncil.com/sf-2026
β‘ FIND US:
X: https://x.com/AICouncilConf
LinkedIn: linkedin.com/company/aicouncilconf
Website: aicouncil.com

![The Deconstructed Database and the Advent of the Open Data Lake
[2025 - Day 3 - Keynote] Julien Le Dem shares insights from his years of experience building the deconstructed database vision, exploring how the big data ecosystem has evolved into a composable architecture organized around open source standards. This talk offers valuable guidance on leveraging interoperable components like Parquet, Arrow, and Iceberg to avoid vendor lock-in, particularly relevant for architects designing data systems or engineers building specialized query engines.
ABOUT THE SPEAKER:
Julien Le Dem, Principal Engineer, Datadog -
ποΈ GET YOUR TICKET TO AI COUNCIL 2026 ποΈ
Meet the worlds top AI infrastructure minds where architects of AI share what works. Three days of high-quality technical talks and meaningful interactions.
β https://aicouncil.com/sf-2026
β‘ FIND US:
X: https://x.com/AICouncilConf
LinkedIn: https://www.linkedin.com/company/aicouncilconf/
Website: https://aicouncil.com/ The Deconstructed Database and the Advent of the Open Data Lake](https://i.ytimg.com/vi/CqhkP2-0Js4/mqdefault.jpg)
![The Future of Data Engineering in the Agent Era
[2025 - Day 1 - Data Eng & Infra] Diptanu Gon Choudhury, Gleb Mezhanskiy, Nuno Campos, and Madison Faulkner share insights from revolutionizing data engineering through AI agents and foundation models, exploring advanced real-time processing and knowledge representation systems. This expert panel offers valuable perspectives on the evolution from traditional data modeling to dynamic knowledge graphs and graph RAG systems, particularly relevant for teams building context-aware data platforms or implementing agent memory systems.
ABOUT THE SPEAKERS:
Diptanu Gon Choudhury, Founder & CEO, Tensorlake
Gleb Mezhanskiy, Co-Founder & CEO, Datafold
Nuno Campos, Founding Engineer, Langchain
Madison Faulkner, Principal & Head of Data Science, NEA
ποΈ GET YOUR TICKET TO AI COUNCIL 2026 ποΈ
Meet the worlds top AI infrastructure minds where architects of AI share what works. Three days of high-quality technical talks and meaningful interactions.
β https://aicouncil.com/sf-2026
β‘ FIND US:
X: https://x.com/AICouncilConf
LinkedIn: https://www.linkedin.com/company/aicouncilconf/
Website: https://aicouncil.com/ The Future of Data Engineering in the Agent Era](https://i.ytimg.com/vi/D3fjQf_OiFg/mqdefault.jpg)
![Building Blocks: Advanced Semantic Data Model Layering
[2025 - Day 2 - Analytics & BI] Lloyd Tabb shares insights from advanced semantic data modeling, exploring how to create larger, more interesting building blocks by reusing queries themselves beyond traditional tables and measures. For data practitioners seeking deeper understanding, this talk offers valuable perspectives on how more sophisticated modeling layers can build comprehension for both humans and AI, drawing from the creator of Malloy and former CTO of Looker.
ABOUT THE SPEAKER:
Lloyd Tabb, Founder & Former CTO, Looker, Malloy Co-Creator, Software Engineer, Meta -
ποΈ GET YOUR TICKET TO AI COUNCIL 2026 ποΈ
Meet the worlds top AI infrastructure minds where architects of AI share what works. Three days of high-quality technical talks and meaningful interactions.
β https://aicouncil.com/sf-2026
β‘ FIND US:
X: https://x.com/AICouncilConf
LinkedIn: https://www.linkedin.com/company/aicouncilconf/
Website: https://aicouncil.com/ Building Blocks: Advanced Semantic Data Model Layering](https://i.ytimg.com/vi/DZkXvdzYlVs/mqdefault.jpg)
![How we made Trail of Bits AI-Native (so far) | Trail of Bits
[2026 - DAY 3 - AI SECURITY & SAFETY] Giving employees AI tools doesnt make a company AI native. Trail of Bits founder and CEO Dan Guido explains the difference β and the operating system his 150-person security firm built to actually get there.
Dan covers the psychological biases that make experts resist AI, the system Trail of Bits built to overcome them, and the results: a jump from 15 to 200 bugs found per week on some engagements. He also gets into what AI-discoverable vulnerabilities mean for security as an industry, and why the real bottleneck has shifted from finding bugs to judging which ones actually matter.
SPEAKER:
Dan Guido - Co-founder & CEO, Trail of Bits
π Sign up for our No BS Newsletter to get the latest technical data & AI content: https://aicouncil.com/newsletter
ABOUT AI COUNCIL:
AI Council brings together the brightest minds in data to share industry knowledge, technical architectures and best practices in building cutting edge data & AI systems and tools.
FIND US:
Website: https://aicouncil.com/
LinkedIn: https://www.linkedin.com/company/aicouncilconf/
X: https://x.com/aicouncilconf How we made Trail of Bits AI-Native (so far) | Trail of Bits](https://i.ytimg.com/vi/DmSVF0ZfTGo/mqdefault.jpg)
![Evals 101: Intro to Evals for Engineers | Braintrust
[2026 - DAY 1 - WORKSHOP] Itβs challenging to understand how complex agents are improving over time and they have many ways that they could go wrong. It is important for AI Engineers to incorporate a disciplined evaluation process as a result. In this session, Curtis Galione will walk through the process of evaluating an agent end-to-end with a hands-on example. Feel free to follow along!
SPEAKER:
Curtis Galione - Solutions Engineer, Braintrust
π Sign up for our No BS Newsletter to get the latest technical data & AI content: https://aicouncil.com/newsletter
ABOUT AI COUNCIL:
AI Council brings together the brightest minds in data to share industry knowledge, technical architectures and best practices in building cutting edge data & AI systems and tools.
FIND US:
Website: https://aicouncil.com/
LinkedIn: https://www.linkedin.com/company/aicouncilconf/
X: https://x.com/aicouncilconf Evals 101: Intro to Evals for Engineers | Braintrust](https://i.ytimg.com/vi/Dn3_H2zcvPI/mqdefault.jpg)
![Orchestrating at Scale How Instacart Manages 20M+ Daily Workflows
[2025 - Day 1 - Data Eng & Infra] Anant Agarwal shares insights from orchestrating at scale, exploring how Instacart manages 20M+ daily workflows and the technical challenges of coordinating massive data operations. For engineers building workflow orchestration systems or scaling data pipelines, this talk offers valuable perspectives on managing complex distributed systems that power one of the worlds largest on-demand platforms.
ABOUT THE SPEAKER:
Anant Agarwal, Staff Software Engineer, Instacart -
ποΈ GET YOUR TICKET TO AI COUNCIL 2026 ποΈ
Meet the worlds top AI infrastructure minds where architects of AI share what works. Three days of high-quality technical talks and meaningful interactions.
β https://aicouncil.com/sf-2026
β‘ FIND US:
X: https://x.com/AICouncilConf
LinkedIn: https://www.linkedin.com/company/aicouncilconf/
Website: https://aicouncil.com/ Orchestrating at Scale How Instacart Manages 20M+ Daily Workflows](https://i.ytimg.com/vi/ECN57ZB9xRs/mqdefault.jpg)
![Stop Worrying About Foundation Models β Your Taste, Judgment, and UX Are the Real Moat
[2025 - Day 2 - Foundation Models] Hamel Husain challenges the common misconception that having a powerful foundation model equals having a successful AI product, drawing from his extensive experience building AI systems at companies like GitHub, Airbnb, and DataRobot. This talk examines the crucial gap between model capabilities and product success, exploring the engineering, user experience, and business considerations that determine whether AI applications actually solve real problems for users.
ABOUT THE SPEAKER: Hamel Husain, ML Consultant, Parlance Labs; R&D Team, AnswerAI
ποΈ GET YOUR TICKET TO AI COUNCIL 2026 ποΈ
Meet the worlds top AI infrastructure minds where architects of AI share what works. Three days of high-quality technical talks and meaningful interactions.
β https://aicouncil.com/sf-2026
β‘ FIND US:
X: https://x.com/AICouncilConf
LinkedIn: https://www.linkedin.com/company/aicouncilconf/
Website: https://aicouncil.com/ Stop Worrying About Foundation Models β Your Taste, Judgment, and UX Are the Real Moat](https://i.ytimg.com/vi/EEw2PpL-_NM/mqdefault.jpg)
![RAG in 2025: State of the Art and the Road Forward
[2025 - Day 1 - AI Engineering] Tengyu Ma shares insights from advancing Retrieval-Augmented Generation beyond current brittleness, exploring how improvements in embedding models and rerankers are creating more robust, scalable solutions. For engineers working with RAG pipelines, this talk offers valuable perspectives on moving away from heuristic-heavy approaches toward automated solutions that deliver higher-quality, more relevant results with less manual optimization.
ABOUT THE SPEAKER:
Tengyu Ma, Co-Founder & CEO, Voyage AI -
ποΈ GET YOUR TICKET TO AI COUNCIL 2026 ποΈ
Meet the worlds top AI infrastructure minds where architects of AI share what works. Three days of high-quality technical talks and meaningful interactions.
β https://aicouncil.com/sf-2026
β‘ FIND US:
X: https://x.com/AICouncilConf
LinkedIn: https://www.linkedin.com/company/aicouncilconf/
Website: https://aicouncil.com/ RAG in 2025: State of the Art and the Road Forward](https://i.ytimg.com/vi/EUxkKELGChM/mqdefault.jpg)
![Scaling CDC to Trillions of Rows: What Broke, What We Rebuilt, and What AI Demands Next | Artie
[2026 - DAY 1 - DATA ENG & DATABASES] Most CDC pipelines work fine when youre building an MVP. Ours did too - until they didnt. Artie is a real-time data replication platform that processes 20-30 billion events per day across thousands of pipelines with sub-minute latency on 90% of them. Three years ago we were running a forked version of Debezium with Kafka processing millions of rows. Along the way, many assumptions we started with broke.
This talk is a post-mortem of what failed, what we rebuilt, and the decisions that matter at scale:
β’ Why we replaced Debezium - single-threaded capture, limited extensibility, and no built-in recovery forced us to build a proprietary Reader from scratch to increase fault tolerance
β’ Parallel backfills without data loss - running historical loads alongside live CDC using primary-key range chunking and exactly-once merge semantics, following Netflixs DBLog pattern
β’ Fan-in from thousands of single-tenant databases - consolidating sharded or single-tenant sources into unified destination schemas without bespoke ETL per tenant
β’ Edge cases at scale - five-digit-year timestamps, negative years, non-JSON in JSONB, non-UTF8 encodings, and why we chose to fail hard rather than silently skip data (and the recovery mechanisms that make that practical in production)
β’ Schema evolution - automatic column adds, type changes, drops, and notifications so teams know what changed
Finally: AI workloads have the same freshness problem databases have always had, but the sources are no longer just databases - they are filesystems, object stores, git repos, and documents. We will share how Artie is extending its core primitives beyond databases to become the sync layer for any data AI systems depend on.
Viewers will leave with concrete architectural patterns for building CDC systems that survive at scale, a checklist of failure modes, and a framework for thinking about real-time data as AI infrastructure.
SPEAKER:
Robin Tang - Co-founder & CTO, Artie
π Sign up for our No BS Newsletter to get the latest technical data & AI content: https://aicouncil.com/newsletter
ABOUT AI COUNCIL:
AI Council brings together the brightest minds in data to share industry knowledge, technical architectures and best practices in building cutting edge data & AI systems and tools.
FIND US:
Website: https://aicouncil.com/
LinkedIn: https://www.linkedin.com/company/aicouncilconf/
X: https://x.com/aicouncilconf Scaling CDC to Trillions of Rows: What Broke, What We Rebuilt, and What AI Demands Next | Artie](https://i.ytimg.com/vi/EcYB5qraEZc/mqdefault.jpg)
![How Action Bias Breaks Autonomous Software Maintenance | LogicStar.ai
[2026 - DAY 3 - LIGHTNING TALK] Coding agents are increasingly trusted to resolve issues end-to-end: investigate, patch, ship, without a human in the loop. But in real-world maintenance tasks, a large fraction of incoming bug reports describe issues that are already fixed. A competent maintainer moves on. Current agents dont. In our new benchmark FixedBench, frontier models apply unnecessary edits to already-correct code in 35-65% of cases, even with full git history and a working environment. More reasoning doesnt help. Better prompts help, but trade one failure mode for another. Todays training rewards producing patches, not deciding whether one is needed. At scale, that quietly compounds into technical debt. The fix starts with framing inaction as a valid success state.
SPEAKER:
Mark Niklas Mueller - Co-founder & CTO, LogicStar.ai
π Sign up for our No BS Newsletter to get the latest technical data & AI content: https://aicouncil.com/newsletter
ABOUT AI COUNCIL:
AI Council brings together the brightest minds in data to share industry knowledge, technical architectures and best practices in building cutting edge data & AI systems and tools.
FIND US:
Website: https://aicouncil.com/
LinkedIn: https://www.linkedin.com/company/aicouncilconf/
X: https://x.com/aicouncilconf How Action Bias Breaks Autonomous Software Maintenance | LogicStar.ai](https://i.ytimg.com/vi/Eqqd7kGD1BM/mqdefault.jpg)