Hire Tanay Shah —
founding AI engineer in NYC.

Available for hire — founding-engineer roles in New York City, May 2026. Most recent: Founding Engineer at Structured AI (Y Combinator F25, NYC). 14-week tenure, 1,035 commits across backend, frontend, iOS, Revit, and infra. I architected the production document-analysis agent platform, built a Zero-Data-Retention Python execution sandbox (bubblewrap + seccomp + AST validation + JSON-RPC over child subprocess), designed a 196-task evaluation harness, and shipped the iOS / iPadOS client from zero in two weeks.

Now in market for the next founding-engineer role at an applied-AI startup or an infra / platform team where the work touches production agent systems.

• → Agent platforms. LangGraph + Anthropic SDK with on-demand tool loading, prompt caching, persistent skill / memory injection. Replaced a 1,448-line single-file agent with a typed-tool runtime.
• → Hardened Python sandboxes. JSON-RPC over child subprocess; AST validator; bwrap + seccomp guards; gRPC + Protobuf bridge to a slim agent pod. ZDR-grade.
• → Eval harnesses. A 196-task benchmark with Redis-locked wave scheduling. Direct Anthropic SDK lean agents, six tools, zero framework overhead.
• → Durable event-sourced backends. Auto-migration container-startup, parallelized staging builds, JSON logging with rotation, DLQ for the durable chat layer. Event-sourced chat replays on reconnect.
• → iOS apps from zero. Apple Sign-In OAuth with Redis fallback for ASWebAuthenticationSession token loss; MarkdownUI streaming chat; 3-column iPad workspace; custom animations and gesture-based navigation.
• → Real-time data pipelines. 8-service market- research platform on AWS Fargate + FastAPI + Redis, enriching and classifying 500K+ companies daily; 20K+ events / hour into live KPI panels with sub-second update latency.

Founders, hiring engineers, and recruiters at applied-AI startups in New York City. Particularly excited about teams working on:

• → Agent infrastructure, MCP servers, multi-vendor LLM orchestration
• → AI safety / runtime hardening / sandboxed code execution
• → Durable event-sourced chat, real-time data, RAG at scale
• → Vertical AI for regulated domains (legal, healthcare, AEC, finance)
• → Developer-tools companies shipping AI products

Email tanayshah2024@gmail.com. I respond within 24 hours.

Public engineering write-ups. Each entry shows the why, the architecture, and the lessons.

• → Structured AI — Founding Engineer field report
• → Travel MCP Server — flights / hotels / weather as tools
• → Mercury Stream — real-time market data with on-anomaly flight recorder
• → AI Agent Error-Handling Patterns — production reliability on Trigger.dev v4
• → Real-Time Sales-Coaching Agent — sub-2s live coach + post-call deep dive

Field-report blog posts on agent infrastructure, sandboxing, MCP, and zero-data-retention design.

• 📝 What the Bubblewrap Sandbox Escape Tells Us About Agent Runtime Hardening in 2026
• 📝 Picking MCP Servers for an Agent — A Selection Heuristic for 2026
• 📝 Building a Zero-Data-Retention Layer for Production LLM Agents

Python · FastAPI · asyncpg · async workers · TypeScript · Swift / SwiftUI · gRPC + Protobuf · PostgreSQL · Redis · Docker multi-stage · Azure Container Apps · AWS (Lambda / Fargate / S3 / SQS) · Kafka · LangGraph · Anthropic SDK · MCP · Next.js · pgvector · Trigger.dev · Apache AGE.

Comfortable across the full agent stack — from sandboxed runtime up through gRPC bridge, async backend, durable chat layer, and iOS client.