Onkar Mahamuni

Hello, I'm

Onkar Mahamuni

Product Software Engineer · Pune, India

I build full-stack products end to end, from .NET and C# backends and data pipelines to React and TypeScript interfaces and AI-powered features.

See my workGet in touch
  • 5,000+
    Load tests ingested
  • 30+ min → 3–4 min
    AI report generation
  • 1st
    Prize, internal AI hackathon, Leading the team

About

Who I am

I am a Product Software Engineer based in Pune. For the past three years I have built and run internal products used across the organization, owning them from architecture through deployment.

The performance intelligence platform I built ingests results from common load-testing and infrastructure monitoring tools, normalizes them into one model, and turns them into time-aggregated, comparable data for dashboards and AI-generated summaries. I also built an AI assistant for Microsoft Teams that sits on top of it.

I care most about the step from messy data to something a person can act on. That is why I spend time on pipelines, storage choices and the performance of the system itself.

Career

Experience

  1. Product Software Engineer

    May 2023 – Present

    Enterprise software company, Pune

    • Architected and built a performance intelligence platform (.NET 10, C#, React, TypeScript) from scratch. It imports results from JMeter, LoadRunner and Azure DevOps, plus PerfMon, DMV and IIS infrastructure data.
    • Designed a multi-store data layer: SQL Server for metadata and thresholds, ClickHouse and InfluxDB for time-series data, Elasticsearch for logs.
    • Built an asynchronous import pipeline on Temporal that parses, aggregates, detects steady states, correlates infrastructure data and compares runs against past results.
    • Redesigned the AI report generator. Batched data is sent to the LLM as parallel async requests and the report is assembled from structured JSON. Generation dropped from 30+ minutes, often timing out, to 3–4 minutes, and hallucinations were largely eliminated.
    • Diagnosed a memory bottleneck where multi-GB raw CSVs were held in C# memory (10+ GB of RAM). Started moving heavy analytics to DuckDB, which is in progress. Ran the database evaluation that led to ClickHouse for high-cardinality load data.
    • A Teams chatbot connected to the performance intelligence platform through MCP servers and RAG pipelines. It can trigger tests, summarize existing tests, analyze results and predict product health, and it delivers user-specific notifications from the platform.
    • Built Grafana dashboards and comparison views, plus an AI summary and natural-language query experience backed by MCP servers and a RAG index.
    • Trained and mentored a group of engineers in C# and led their knowledge transfer.

  2. Individual Contributor, Data Visualization Platform

    2024

    Enterprise software company, Pune

    • Built a tool that takes Excel, CSV or SQL data, infers its structure and generates charts on demand. Placed runner-up at the global round of an internal competition.

Selected work

Projects

Internal hackathon 2026 · 1st Prize · Team lead of 6 engineers

Agentic SDLC Performance System

An agentic AI system with an agent at each SDLC stage, from requirements to production monitoring. Each agent reviews its stage for performance, and a central agent shares incidents and learnings through a graph database. When tests fail, the system finds the root cause and, for code issues, raises a fix PR and re-tests once it is approved.

  • LangGraph
  • FastAPI
  • React
  • MCP servers
  • JMeter
  • Graph database

Internal AI hackathon · 3rd Prize

Multi-Agent VM Right-Sizing Agent

Led a team building an AI agent that forecasts load from seasonal usage patterns (Prophet), correlates user counts with Datadog metrics, and recommends the lowest-cost Azure VM configuration. Scheduled scale-ups and scale-downs need human approval and then run through the Azure APIs.

  • Python
  • LangGraph
  • Prophet
  • Azure APIs
  • React

2023 – Present

Performance Intelligence Platform

An internal data platform I built and run. It ingests results from several sources, normalizes and aggregates them over time, and feeds Grafana dashboards, comparison views and AI summaries.

  • .NET 10
  • C#
  • React
  • Temporal
  • ClickHouse
  • DuckDB
  • MCP Servers
  • AI Agents
  • Grafana

AI assistant · Microsoft Teams

AI Performance Assistant

A Teams chatbot connected to the performance intelligence platform through MCP servers and RAG pipelines. It can trigger tests, summarize existing tests, analyze results and predict product health, and it delivers user-specific notifications from the platform.

  • MCP servers
  • RAG pipelines
  • Microsoft Teams
  • LLM agents
  • Notifications

2024 · Internal competition, runner-up (global round)

Data Visualizer

Upload an Excel or CSV file, or connect a SQL database. The tool works out the structure of the data and produces the visualizations the user asks for.

  • Data visualization
  • Automatic schema detection

Recognition

Awards

1st Prize

Internal hackathon, 2026

Led a team of 6 engineers

3rd Place, Global

Global final of the internal tournament, 2026

Live product demo to company CXOs, as team lead

3rd Prize

Internal hackathon, 2025

Led a team building a multi-agent cloud cost and autoscaling agent

Runner-up, Global Round

Internal competition, 2024

Data visualization platform

Toolbox

Skills

Backend

  • C#
  • .NET 10
  • ASP.NET Core
  • Worker services
  • Clean Architecture
  • CQRS-lite
  • Dapper
  • Temporal
  • RabbitMQ

Frontend

  • React
  • TypeScript
  • TanStack Query
  • TanStack Table
  • TanStack Router
  • React Hook Form
  • Zod
  • Azure MSAL

Data

  • SQL Server
  • ClickHouse
  • PostgreSQL
  • TimescaleDB
  • InfluxDB
  • DuckDB
  • Elasticsearch
  • Grafana

AI & Automation

  • LLM integration
  • RAG pipelines
  • MCP servers
  • Agent orchestration
  • LangGraph
  • Python

Cloud & Ops

  • Azure VMs
  • Entra ID
  • Azure DevOps
  • Datadog

Background

Education

PG Diploma in Advanced Computing

CDAC Sunbeam, Pune

Sep 2022 – Mar 2023

C-CAT Aug 2022: AIR 136

Bachelor of Engineering, Mechanical

SKNSITS, Lonavala · University of Pune

2015 – 2019

Contact

Let's talk

I'm open to conversations about full-stack product development, backend and data platforms, and AI-assisted tooling.

Email meLinkedInGitHub