Open to opportunities

Rajat Balyan

@rajatbalyan

Applied AI engineer building production LLM fine-tuning and multi-cloud RAG systems that reduce latency and hallucinations.

India

Message

What I'm looking for

I’m seeking a dedicated full-time remote Applied AI role, focused on production engineering—LLM fine-tuning, multi-cloud RAG, and systems that reduce latency and hallucinations, with a strong bias toward shipping reliable features.

I’m an applied AI engineer with production experience fine-tuning 671B-parameter LLMs, building multi-cloud RAG pipelines, and shipping autonomous developer tooling—built for real deployment environments.

As the lead AI engineer, I architected a GCP-native legal AI platform that routes live queries against a 100M-row Parquet dataset on AWS S3. I implemented a hybrid BM25 + semantic retrieval pipeline, dynamic subset selection, reranking, citation grounding, and cross-cloud latency management between GCP and AWS.

I also led a layered hallucination-mitigation strategy—hard prompt constraints, live retrieval integration, structured citation outputs, and post-generation citation verification—addressing a production degradation incident. I scoped and reviewed contractor architecture and redesigned key gaps with an async UX and phased rollout plan.

I’ve been recognized with 1st Prize at IIT Roorkee Cognizance 2025 for an autonomous CI-driven maintenance platform that reduced weekly upkeep by ~70%, and I secured $25K in GCP credits while holding 69 Google Cloud Skill Badges. I’m focused on full-time remote applied AI/ML work where engineering results matter.

Experience

Work history, roles, and key accomplishments

Current

Lead AI Engineer

Current

Asvara Innovations Pvt. Ltd.

Jan 2024 - Present (2 years 5 months)

Architected PleadSmart’s multi-cloud (GCP↔AWS) RAG pipeline for an AI legal chat, routing queries over a 100M-row Parquet dataset in AWS S3 using hybrid BM25 + semantic retrieval with reranking and verified citations. Implemented layered hallucination mitigation to resolve a production DeepSeek R1 degradation issue, and delivered Site Sentry automation that cut weekly upkeep by ~70% while securing

RAG Hybrid Retrieval (BM25 + Semantic Search)Faiss Cloud Functions GCP AWS Cross Cloud Routing

Current

AI Systems Engineer

Current

Meta Catalyst Pvt. Ltd.

Jan 2023 - Present (3 years 5 months)

Led product and technical direction for an AI + IoT automation company, focusing on software/task automation agents and planning expansion toward physical robotics systems. Explored multi-agent frameworks (Fetch.ai/uAgents) and Web3 + AI integrations for autonomous agent workflows.

Agentic Systems Fetch.Ai UAgents AI Agents IoT Sensor Integration Software Task Automation Arduino Autonomous Robotics Planning