About

AI/ML Engineer building production LLM applications with Python: multi-agent systems, RAG pipelines and real-time voice agents. Shipped a low-latency voice agent and multimodal enterprise RAG using LangChain, LangGraph and FastAPI, with automated evaluation, tracing and deployments on AWS.

Experience3

1. EG

   AI/ML Engineer Intern

   Jan 2026 — Aug 2026

   Etech Global Services · Remote

   - Engineered a real-time voice agent (LiveKit) that resolves customer calls over phone and web, achieving a p95 time-to-first-audio of 1.2 seconds. - Delivered a multimodal RAG pipeline that improved retrieval relevance by 40% through structure-aware chunking. - Enhanced grammar-correction accuracy from 40% to over 90% by developing a hybrid pipeline of NLP pre-filtering and LLM validation.

   • Achieved 1.2s p95 time-to-first-audio for voice agent
   • Improved retrieval relevance by 40%
   • Increased grammar-correction accuracy from 40% to 90%+
2. II

   Research Intern, AI in Healthcare

   Jan 2026 — Jun 2026

   Indian Institute of Technology (IIT) Ropar · Rupnagar, Punjab

   Curated 918 raw CTA scans for aneurysm detection, successfully converting them via DICOM and extracting vessels. Fixed coordinate mapping for approximately 50% of positive patches, resulting in 1,971 patches with stratified 5-fold splits.

   • Curated 918 raw CTA scans to 518 for aneurysm detection
   • Yielded 1,971 patches with stratified 5-fold splits
3. P(

   Machine Learning Intern

   Feb 2025 — Mar 2025

   PW (PhysicsWallah) · Remote

   Built a LightGBM model predicting revenue from over 900K e-commerce user sessions, achieving an RMSE of 1.1. Developed a Streamlit and Plotly dashboard for RFM segmentation and traffic KPIs, serving predictions in 15 milliseconds.

   • Achieved 1.1 RMSE for revenue prediction model
   • Served predictions in 15 ms with dashboard

Projects2

• AI Sales Orchestrator (Multi-Agent System)

  Architected a stateful multi-agent sales system for product discovery, checkout, delivery and support, coordinating 6 specialized agents and preserving cart context across platforms.

  Open on github.com

  • Python
  • LangGraph
  • FastAPI
  • AWS

• Enterprise RAG (Document Q&A)

  Built an agentic pipeline that routes questions to document search, web search or direct chat, achieving high contextual recall and precision.

  Open on github.com

  • Python
  • LangGraph
  • FastAPI
  • Weaviate
  • Redis
  • LiteLLM

Education1

• KU

  B.Tech in Computer Science and Engineering (AI and ML)

  2022 — 2026

  Kalinga University

Certifications1

• Advanced Certification in Data Science and AI

  iHUB DivyaSampark, IIT Roorkee, with Intellipaat