Looking for a job

Swayam Yadav

@swayamyadav

Full-stack developer building production AI products—LLM pipelines, secure backends, and polished UIs with rapid shipping.

India

Message

What I'm looking for

I’m looking for a role where I can own production AI systems end-to-end—LLM document pipelines, secure APIs, and clean frontends—shipping quickly with real customer impact in a fast-moving team.

I’m a full-stack developer with production experience across NYC-based AI and legal-tech startups, focused on building end-to-end systems powered by LLMs. I built an LLM-powered document pipeline processing 100+ documents weekly, and I single-handedly shipped a CLM platform backend (28 REST API endpoints) in under 4 weeks.

On the backend, I designed a JWT dual-token auth system (15-min access / 7-day rotating refresh) with RBAC, implemented multi-format ingestion (DOCX, PDF, OCR → OpenAI gpt-4o-mini → Zod validation → PostgreSQL), and delivered clause intelligence like risk scoring and Present/Weak/Missing clause mapping. As a founder, I built ArcForge (live) solo from zero-to-live and resolved critical post-launch production issues without downtime, and I also contributed and merged a Neo4j toolkit to the agno-agi/agno open-source AI framework.

Experience

Work history, roles, and key accomplishments

Software Developer

Vidhi Setu

Apr 2026 - May 2026 (1 month)

Sole-developed a legal-tech CLM backend, shipping 28 REST API endpoints in under 4 weeks and implementing JWT dual-token auth with RBAC. Built a multi-format document ingestion pipeline (DOCX/PDF/OCR → OpenAI gpt-4o-mini → Zod validation → PostgreSQL), added clause intelligence (risk scoring, clause mapping, Word export), and deployed to Hostinger VPS (Nginx/PM2).

TypeScript Bun PostgreSQL Prisma JWT RBAC OCR Document Ingestion OpenAI API Zod Validation

Software Development Intern

SkillRank

Oct 2024 - May 2025 (7 months)

Built the primary FastAPI backend for CivicSight AI, processing 100+ documents weekly at ~95% extraction accuracy. Designed an AGNO-based multi-agent workflow, contributed a Neo4j toolkit merged into agno-agi/agno, and improved reliability and performance by rebuilding AWS Lambda triggers with idempotent retry logic and adding LRU caching to cut query latency from ~850ms to under 150ms.

fastAPI Python Agno Neo4j AWS Lambda LRU Caching Performance Optimization