Open to opportunities

Vatsal Agrawal

@vatsalagrawal

Senior Backend Engineer with ~6 years building production systems in Python and Golang, now focused on shipping LLM-powered products end-to-end.

India

Message

What I'm looking for

I’m looking to own end-to-end backend and GenAI system delivery—voicebots, RAG pipelines, and performance tuning. I want a team where I can lead architecture, ship reliably in production, and keep tail latency and SLAs under control.

I’m a Senior Backend Engineer with 5+ years across Python and Golang, focused on diagnosing performance bottlenecks in production systems and building production GenAI systems (voicebots, RAG pipelines with hybrid search and LLM reranking). I’m comfortable owning systems end-to-end and leading small teams through architecture decisions and code review.

Most recently, at Omniful, I led a team of 5 engineers and shipped a PIM module that cut seller onboarding time by ~60%. I also collapsed 30 MongoDB indexes into 5, bringing API tail latency from 5s to 75ms, and built an audit-grade Action Log on Kafka for model-change tracking at 50,000+ events/day. Previously at Squadstack, I delivered a production voicebot with ~80% containment, resolved PostgreSQL hot paths with targeted Redis caching, and built an agent-fulfilment engine that hit SLAs on ~95% of 50,000 daily leads.

Experience

Work history, roles, and key accomplishments

Software Development Engineer 2

Omniful

Aug 2025 - Apr 2026 (8 months)

Led a team of 5 engineers, owning architecture decisions and cross-service code reviews. Built a PIM module that reduced seller onboarding time by ~60%, migrated MongoDB to Atlas Search by collapsing 30 indexes into 5, and designed a Kafka action log scaling to 50,000+ model-change events/day.

Golang MongoDB Atlas Kafka Microservices RESTful APIs PostgreSQL Redis CI CD Amazon SQS

Product Engineer (PE2)

Squadstack

Jul 2020 - Jul 2025 (5 years)

Shipped a production regional-language voicebot handling ~600 daily queries at ~80% containment, later extending it with a RAG-powered voicebot for complex e-commerce support. Resolved a PostgreSQL hot path by restructuring queries and adding Redis caching to reduce p90 from 5s to 75ms, and built an agent-fulfilment engine achieving ~95% SLA attainment across 50,000 leads/day.