Open to opportunities

Rohit Bind

@rohitbind

Message

GenAI engineer building fast, production LLM systems with agents, caching, and observability.

India

Message

What I'm looking for

I’m looking for a role where I can build production LLM/agent systems end-to-end—optimizing latency and reliability, adding observability, and iterating on retrieval and evaluation to deliver measurable impact.

I’m an AI engineer who turns LLM ideas into reliable, production-grade systems—focusing on agentic workflows, latency control, and measurable quality. At Swarmlens, I shipped a multi-round ReAct agent loop with 13 custom tools and search grounding for production AI chat.

I’m hands-on with performance and failure modes: I built a Redis cache-aside layer (326× speedup), designed a 3-tier LLM fallback chain with rollback recovery, and reduced end-to-end response latency 80% by tuning function rounds and caching enriched search results. I also implemented per-turn observability persisted in JSONB to support SQL P95 latency analytics across production conversations.

Before that, I helped improve retrieval and cost efficiency as an AI Engineer Intern at Satsure.co—vectorizing 400K+ records, deploying a GPU Weaviate cluster, and raising top-k precision via multi-stage re-ranking while cutting token costs 40% using snippet ranking and a CrewAI multi-agent pipeline.

Experience

Work history, roles, and key accomplishments

Founding GenAI Engineer

Swarmlens

Dec 2025 - Mar 2026 (3 months)

Engineered a multi-round ReAct agent loop with 13 custom tools and Google Search grounding for production AI chat. Improved performance by 326× using a Redis cache-aside layer and reduced end-to-end response latency 80% (70s → 14s) via LLM fallback tuning, prompt compression, and SSE streaming delivery.

Function Tool Orchestration Grounding LLM Fallback Chains SSE GraphRAG With Spanner Graph

AI Engineer Intern

Satsure.co

Jun 2025 - Nov 2025 (5 months)

Built a multi-stage re-ranking pipeline (BM25 + Vector + RRF + Cross-Encoder), improving top-k precision by 24%. Vectorized 400K+ records, launched a GPU Weaviate cluster to cut query latency from hours to minutes, and reduced LLM token costs 40% by preventing context overflow.

BM25 Vector Search RRF Re Ranking Cross Encoder Re Ranking Weaviate Snippet Ranking

Data Analyst Intern

Vending Brothers

Jun 2024 - Aug 2024 (2 months)

Analyzed purchase patterns across 210 vending machines, reducing refill trips by 20%. Identified product demand across 50 locations, increasing sales by 12% and improving profit margins.

Location Based Demand Analysis Data Driven Decision Making

Operations Mgmt Intern

Indian Oil Corporation Limited

May 2023 - Jul 2023 (2 months)

Performed CFD modeling using FVM, reducing power usage and drag while improving efficiency by 3%. Built SCADA analytics-based pump defect prediction to reduce unplanned downtime across refinery operations.

Predictive Maintenance Pump Defect Prediction Efficiency Improvement