AI Engineer NYC 2026
as the organisers put it
Where AI engineering meets financial services. Join 1,000+ engineers and finance leaders at the Sheraton New York Times Square, October 12-14, 2026.
who it's for
ai engineers, ml engineers, security engineers, platform engineers, site reliability engineers, engineering managers
topics
self-hosted open models deployment; vLLM benchmarking and configs; agent identity and authentication; agent sandboxing and containment; automated code review agents; AI-driven logic bug hunting; supply-chain and agent security; merge automation at scale; durable execution for agents; quantization and tensor-parallel tuning; repo-to-component software graph; agent governance and lifecycle
| title | author | length |
|---|---|---|
| Catch the Bugs Your Code Review Misses | Killian Carlsen-Phelan, Taylor Luttrell-Williams | 38m |
Gitar is set up against a sample Box Office ticketing app and runs three staged PRs: it finds a bug outside the diff where a buyer can refund a ticket after… | ||
| Why Humans Can't Keep Up With AI-Written Code | Tariq Shaukat | 17m |
Agents produce 2,450% more lines of code but releases rise only ~30%, and five failure modes matter: humans agree with wrong AI 4 out of 5 times, agents fail… | ||
| How Revolut Cut AI Coding Costs 90% With Self-Hosted Open Models | David Abitbol | 12m |
Self-hosted GLM deployment cuts inference spend roughly 10x vs Anthropic/OpenAI and claims about 90% cost reduction with a blended token cost of 30¢ per… | ||
| We Trust AI Agents That Break Every Day | Shy Ruparel | 22m |
Survey finds 80.8% daily use of agents but only 49.1% in production, with 85.5% saying they trust output while 41.1% see issues daily and 22% lose progress to… | ||
| Why a One-Line Change Can Break Your Whole Stack | Cole Gannaway | 15m |
A one-line data model change in Bitwarden passes tests and linters but silently breaks web, mobile, CLI and MCP clients across repos, showing how small edits… | ||
| How Anthropic Merges 30,000 Code Changes a Day | Manish Kapur, Addy Osmani | 19m |
Anthropic reports scaling from about 2,500 to 30,000 merges a day by using Claude first-pass review that produces runnable verification evidence (end-to-end… | ||
| AI That Hunts Logic Bugs Before Attackers Do | Taylor Luttrell-Williams | 14m |
SonarQube Hunter Agent maps a repo, runs three specialist agents, adversarially validates each lead and exports findings into SonarQube. A one-hour demo on a… | ||
| Stop Guessing Your vLLM Config: Benchmarking LLMs on Runpod | Chris Vela | 16m |
BenchPod is a vLLM-based benchmarking dashboard that lets teams pick configs from data instead of guessing, showing a live demo of BenchPod, running Kimi K3 on… | ||
| Who Is Making This Request? Agent Auth 101 | Kamil Potrec | 19m |
Agents must present per-agent identity and use short-lived scoped credentials with delegation, policy and audit; OAuth, workload identity federation or PKI can… | ||
| Exploits Now Beat Patches: Open Source Security in the AI Era | Patrick Smyth | 22m |
Exploit timelines invert: Mandiant shows exploits hit on average -70 days before patches, Axios backdoor runs two hours and 600,000 downloads, a 17‑year‑old… | ||
| We Kept Rebuilding Our AI Agent. Here's What Survived | Fuad Ali | 20m |
An in-product agent called Alex evolves from a 2023 prototype with ~50% hallucination to a single agent by June 2026 using only custom harness code (no agent… | ||
| Stop Giving Your Agent Tools. Give It a Sandbox | Mikyo King | 15m |
Code mode replaces giant tool catalogs with a programmable sandbox that searches and composes tools as code, running Bash, a Python sandbox, SQL and… | ||
| Why Are We Still Governing AI Agents Like Software? | Tim Tyler | 18m |
Details of the Agent Accountability framework map exactly to what gets built: a registry that issues a single unique ID per agent, assignment into teams and a… | ||
| 300 Agents, 60 People: How We Secured Our Own AI Agents | Aiden Sochowski | 17m |
Runlayer deploys a gateway and logs every session and tool call across 300 agents for 60 people and discovers concrete failures: prompt injection through email… | ||
| The Highest Loop | swyx | 07m |
Claims that productivity comes from climbing and descending nested execution loops and that conferences are the loop that makes loops, organizing engineers to… | ||
| AI Engineer NYC: OpenAI, Blackstone, Ramp, Coinbase, Apollo, Citadel, Basis, Runway, Compute Trading | Ksenia Se, Alexander Shevchenko, Tariq Shaukat, Evan Kormos, Joseph Kim, Philipp Schmid, Joao Wenzel, Will Brown, Jemima Abu, Jiachen Jiang | 08h 52m |
Three days of applied AI for finance land on concrete moves not hype: firms report post-training specialist models (Quen 3.5 35B) to speed spreadsheet Q&A, a… | ||
| AI Engineer NYC Livestream | Bridgewater, Two Sigma, Arize, Coatue, Nous (Hermes), Gen. Intuition | Lia McBride, Alex Lieberman, Aaron Linsky, Sam Green, John Feminella, Akshat Bubna, Rayan Krishnan, Jack Woodruff, Vinoo Ganesh, David Abitbol, Sid Pednekar, Frank Long, Pim de Witte | 07h 38m |
Live from AI Engineer NYC, sessions give concrete outcomes not vaporware: Bridgewater explains a production vault indexing 100 million documents and how… | ||
| Building AIE: Open Q&A and Feedback Session | swyx | 59m |
AIE explains concrete decisions and pain points: it selects speakers from 2,200 applications for ~300 spots, will keep the World's Fair at Moscone next year at… | ||
| nothing matches | ||