Headline

Building AI x Growth Systems · LLM agents, RAG & Applied AI · GTM

About

I'm 21. Shipped production AI at 4 early-stage startups as the sole engineer. Not "contributed to" - shipped. Last year I shipped agentic platforms, RAG pipelines, custom MCP servers, multi-agent systems at scale. Real products. Real on-call moments. All solo. I build with Claude Code like a co-founder - and I write about it publicly. 1,200 technical followers on X and growing. Best engineers understand distribution, not just architecture. Working on both. Stack: LangGraph, LangChain, PostgreSQL, FastAPI, PyTorch, AWS, Pinecone. Looking for founding engineer, AI engineer, devrel, or GTM roles where decisions actually ship. 🔗 shreyaspandey.me

Location

Greater Delhi Area

AI stack

antigravity

codex

claudecode

Work experience

AI Engineer Intern at Cellyn, Feb 2026 - Apr 2026 · 3 mos, Remote, Sole engineer. Built a full enterprise AI platform from zero to production. • Shipped an agentic Excel add-in with LLM tool-calling - analysts can query, model, and reason over live financial data in natural language • Built a multi-tenant secure Data Room: encrypted document storage, room-based RBAC, audit logging, real-time notifications, and AI-powered Q&A on top • Implemented context-aware memory management, sandboxed code execution, automated PPT generation, and web search integration • Owned all cloud infra on Scaleway - CI/CD, deployment workflows, automated testing, production releases Stack: LangGraph · FastAPI · PostgreSQL · Docker · Scaleway · GitHub Actions

AI Engineer Intern at Fills Ai, Nov 2025 - Feb 2026 · 4 mos, Remote, Shipped 4 production systems in 3 months. • Built a semantic RAG pipeline over 30M+ PubMed papers - 92% citation accuracy in scientific contexts. Most teams don't get past 70% • Engineered a real-time collaborative scientific editor (Docs meets AI co-author) - LangChain + Gemini + Tiptap + Y.js with automated version control and cloud sync • Built custom MCP servers for agentic orchestration across separately hosted services • Developed a multi-agent pipeline for end-to-end research automation: literature search, synthesis, and citation-accurate writing Stack: LangChain · Gemini · PGVector · AWS · React.js · Next.js

AI Engineer Intern at Stealth AI Startup, Sep 2025 - Nov 2025 · 3 mos, Remote, Sole AI engineer. Owned the full stack on a real-time voice AI support system - from model integration to production deployment. • Built the end-to-end STT → LLM → TTS pipeline for real-time conversational AI in customer support • Pushed system latency down to 36ms through parallel audio buffering and pipeline optimization (for reference: human reaction time is ~250ms) • Owned all inference infrastructure and DevOps on Scaleway - model serving, real-time audio streaming, scalable deployment • Implemented async processing and parallel audio buffering to handle concurrent voice sessions without degradation Stack: Python · Scaleway · STT/TTS APIs · FastAPI · Docker

AI Engineer Intern at 21 Spheres, Apr 2025 - Sep 2025 · 6 mos, Remote, First real taste of production agentic AI - built multi-agent systems across 5 client deployments. • Built 5 multi-agent ReAct systems handling 1,000+ daily queries with 88% autonomous resolution rate - meaning the agent figured it out without human fallback 9 in 10 times • Architected a multi-agent ReAct framework for autonomous task completion, cutting client processing time by 60% • Led a 3-person engineering team end-to-end - architecture decisions, code reviews, delivery timelines • Designed agent memory, tool-use, and fallback logic for edge cases that would otherwise require human intervention Stack: LangChain · LangGraph · FastAPI · PostgreSQL

Education

Bachelor of Technology, Computer Science, kiet group of institutions

Bachelor of Science, Data Science and Applications, Indian Institute of Technology, Madras

Public links

github: https://github.com/Shreyas-prog108