An AI Engineer who
Turns ideas into
things that work...
Because AI is cool, but useful AI is cooler.
|
Currently, I'm an Associate Software Engineer atZemoso Technologies
I like data. I like building things. And lately, I've been especially interested in what happens when you combine the two with AI. I work with Python, Generative AI, LLMs, RAG, and machine learning to turn ideas into practical applications. Always curious, always experimenting, always building :)
Work Experience
I'm currently looking to join a cross-functional team
that values building practical AI solutions and using technology to solve real-world problems.
Featured Project
AI-Powered Discovery Platform
A GenAI-powered discovery platform that automates client discovery by engaging stakeholders in intelligent, context-aware conversations to surface business requirements and knowledge gaps. Using LLMs, it dynamically generates questions, analyzes responses, and produces structured discovery reports with actionable insights — reducing manual effort and accelerating project planning and solution design.
- Designed and built the Human-in-the-Loop workflow for interview question management — backend APIs, schema changes, and an inline review/approve UI — letting reviewers edit and approve LLM-generated questions before live interviews.
- Built a Knowledge Base ingestion pipeline supporting multiple document formats, with LLM-driven overview generation that auto-produces workspace summaries and 5 contextual discovery questions from uploaded documents.
- Developed the Workspace Detail Page — a full-stack feature covering the interview lifecycle from workspace creation and document upload through candidate invite gating (blocked until questions are approved) to report viewing.
Featured Project
Root Cause Analysis Platform
RCA Accelerator is a GenAI-powered platform that automates root cause analysis for incidents using a multi-agent AI architecture. It builds a knowledge graph linking code commits, deployments, infrastructure telemetry, and incidents, then integrates with tools like GitHub, Jira, Prometheus, and Splunk (via MCP) to combine RAG with reasoning — automatically surfacing root causes, evidence, and confidence-scored RCA reports to speed up incident resolution.
- Built a multi-agent LangGraph pipeline (Orchestrator, Planner, Workers, RCA agents) with retry logic, confidence-gated early stopping, and SSE streaming for real-time investigation.
- Designed a dual-agent RCA system where a primary and conservative analyst independently generate confidence-scored reports, arbitrated by an LLM judge that resolves disagreements and validates citations.
- Engineered a plan-driven MCP integration where the Planner emits structured tool calls, executed by specialized workers (Metrics, Logs, GitHub, Graph) against live observability endpoints to gather evidence.
