AI Engineer with 1.8+ years of experience building LLM agent pipelines and backend services using Python, LangGraph, RAG, and FastAPI. At UBS I've delivered production multi-agent apps with Azure OpenAI and human-in-the-loop approvals. I also build personal projects like AgriMentor and Chikki.
Trainee AI Engineer
UBS
- Request Management Service: Delivered 5–7 enterprise applications onto a LangGraph-based multi-agent foundation, with a reflection agent that reviews outputs and advances the workflow without manual intervention.
- Built onboarding logic that generates SQL commands from agent output, streamlining data setup for each new application.
- Integrated a Microsoft Teams bot for human-in-the-loop approval, routing AI-generated responses to reviewers before final action.
- Deployed and maintained the service in a live environment; recognized by the business team for quality and reliability.
- Provisioning Accelerator: Built the authentication module with secure sign-in using Microsoft Entra ID and Microsoft Graph API.
- Developed file-handling logic to generate and manage files in the provisioning workflow.
- Vibe Testing: Converted BDD specifications into runnable test spec files using GitHub Copilot and validated the AI-generated scripts against application behavior.
- Created NavLearn, a skill-file agent that writes test scripts using a nav-registry and action-registry, cutting manual test-authoring effort by 30–40%.
- Collaborated with QA and engineering teams to integrate NavLearn into the Vibe Testing pipeline.
QA Intern
MassMutual india
- Performed ETL testing using Informatica and Microsoft SQL Server, validating data consistency between on-prem and cloud environments.
- Automated regression scenarios using Selenium and raised defects in JIRA within a Kanban-based Agile setup.
- Performed functional and regression testing for the Eagle Version Release (EVR) against business requirements.
Chikki – Voice-First AI Career Assistant
Voice-first AI portfolio assistant built with React, FastAPI, and WebSockets, using browser speech recognition and Edge-TTS for spoken replies. Ordered multi-model fallback (Groq, Gemini, NVIDIA, OpenRouter) with Langfuse observability for reliable conversations. Google Drive-to-ATS pipeline that generates recruiter-tailored PDF resumes from the conversation.
AgriMentor – AI-Powered Farming Assistant
AI-powered web app giving first-time farmers personalized crop and soil recommendations for Indian conditions, using a RAG pipeline over agricultural knowledge. FastAPI backend APIs that process soil test inputs (PDF/image/text) and return crop recommendations. Streamlit frontend for uploading soil reports and chatting with the assistant. Crop management assistant covering watering schedules, fertilizer timelines, and pest/disease guidance.
B.Tech, Computer Science and Engineering
Kamala Institute of Technology and Science · 7.59 CGPA
Azure AI Fundamentals (AI-901)
Microsoft
Apr 2026Microsoft
Apr 2025Agentic AI Certified Foundations Associate
Oracle
Aug 2026
Attendee · Oct 24, 2026