Open to opportunities

Nikhil Sharma

@nikhilsharma17

Message

I build production-grade agentic AI systems, RAG pipelines, and reliable multi-agent platforms.

India

Message

What I'm looking for

I'm looking to build production-grade agentic AI and GenAI systems where I can own architecture through deployment, with strong evaluation, observability, guardrails, and CI/CD practices for reliable enterprise AI.

I'm building Agentic AI solutions for PepsiCo through Grid Dynamics, including multi-agent Agentic RAG pipelines that improved answer coverage by approximately 35% while maintaining under-30-second end-to-end latency.

I develop orchestration platforms on AWS Bedrock AgentCore using CrewAI, LangGraph, FastAPI, Temporal, MCP, and A2A protocols. My work includes dynamic agent deployment, semantic memory, tool discovery, multi-tenant access control, observability, and asynchronous CI/CD.

Previously at TechStax, I deployed predictive models for PCB manufacturing that reduced raw-material loss by 22%, and built an LLM-powered chatbot using Llama 3 for natural-language database queries.

At Capillary Technologies, I improved retail analytics models for more than 45 clients, reduced campaign costs by 4%, optimized production RFM training time by 11%, and shipped an OpenAI-powered pull-request review application.

Experience

Work history, roles, and key accomplishments

Current

Data Scientist [Agentic AI at Pepsico]

Current

Grid Dynamics

Apr 2025 - Present (1 year 5 months)

Architected and developed a multi-agent Agentic RAG pipeline with LLM-driven query decomposition and iterative self-reflection, improving answer coverage by ~35% while maintaining <30s E2E latency. Built a multi-agent orchestration platform on AWS Bedrock AgentCore enabling dynamic deployment of CrewAI and LangGraph agents via a unified FastAPI REST API, and integrated MCP Gateway and Bedrock Memo