Skip to content
View thebharathkumar's full-sized avatar
✨
In the making
✨
In the making
12 followers · 19 following

Achievements

Achievement: Pair ExtraordinaireAchievement: Pull Sharkx2Achievement: QuickdrawAchievement: YOLO

Achievements

Achievement: Pair ExtraordinaireAchievement: Pull Sharkx2Achievement: QuickdrawAchievement: YOLO

Highlights

Block or report thebharathkumar

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
thebharathkumar/README.md

Hey, I'm Bharath 👋

Typing SVG

Portfolio LinkedIn Medium Twitter Spotify Profile views

banner

💫 About Me

Coding GIF

  • 🎓 MS in Computer Science at Pace University Seidenberg (GPA 3.86, May 2026)
  • 🛠️ Currently building multi-agent chatbot platforms, RAG pipelines, and LLM observability dashboards serving multiple university departments
  • 🤖 Working with Claude on Amazon Bedrock, LangGraph, MCP servers, and production AI infra on AWS & Azure
  • 📚 Published 2 ML research papers (Springer Nature, IJARESM) on computer vision and deep learning
  • 🌱 Going deep on agentic systems, evals, and frontier model tooling
  • ✍️ I write about AI engineering on Medium
  • ⚡ Fun fact: I can solve a Rubik's cube in under 2 minutes

🚀 What I'm Building

🤖 Multi-Agent Chatbot Platform

Production AI assistant platform at Pace Seidenberg using LangGraph orchestration, Claude on Amazon Bedrock, and custom MCP tool servers. Powers multiple university departments with department-specific knowledge bases and routing.

Stack: Python · LangGraph · Bedrock · MCP · FastAPI · AWS

📊 LLM Observability Dashboard

End-to-end observability for production LLM workflows: trace capture, eval harness, cost tracking, and drift monitoring. Built to give faculty and staff visibility into model behavior across the platform.

Stack: TypeScript · React · FastAPI · Postgres · Azure

🔁 Open Source @ Artie Labs

Contributing to artie-labs/transfer, a Go-based real-time database streaming platform. PRs covering CONTRIBUTING docs, an Elasticsearch destination, and a Kinesis source consumer.

Stack: Go · Kafka · Elasticsearch · Kinesis

🌿 Plant Disease ML Pipeline

End-to-end CNN pipeline trained on 50,000+ records for plant disease classification, completed during my research internship at Compsoft Technologies and published with Springer Nature.

Stack: PyTorch · OpenCV · scikit-learn · Python

📚 Research Publications

A Deep CNN Approach for Identifying Medicinal and Edible Plants in the Western Ghats

Atlantis Press

Deep learning approach for identifying medicinal and edible plants from the biodiversity-rich Western Ghats using CNNs, supporting conservation and sustainable use of plant resources. Corresponding author.

Driver Drowsiness Detection using AI

IJARESM

Real-time computer vision system for detecting driver drowsiness using facial landmarks and eye aspect ratio analysis to help prevent fatigue-related accidents and improve road safety.

🛠️ Tech Stack

🤖 AI / LLM Engineering

Claude LangChain LangGraph MCP Bedrock Hugging Face OpenAI

🧠 ML / Deep Learning

PyTorch TensorFlow Scikit-learn Keras OpenCV Pandas NumPy

💻 Languages & Backend

Python Go TypeScript C++ FastAPI Node.js React

☁️ Cloud, MLOps & Infra

AWS Azure GCP Docker Kubernetes Terraform Linux

🗄️ Databases & Data

Postgres MongoDB Redis Elasticsearch Kafka

📜 Certifications

Anthropic AWS SAA RHCSA

📊 GitHub Stats

Activity Graph

🤝 Let's Connect

Portfolio LinkedIn Medium Twitter Instagram

Open to AI Engineer and Software Engineer roles. Currently work-authorized on F-1 OPT.

Pinned Loading

  1. agent-triage agent-triage Public

    Ranks multi-agent failures by severity, frequency, and recovery from OpenTelemetry or NDJSON traces. CLI, FastAPI dashboard, OTLP receiver, optional LLM root-cause analysis. pip-installable, typed,…

    Python

  2. mcp-otel-audit mcp-otel-audit Public

    Public audit of four MCP OpenTelemetry instrumentations against the official OTel semantic conventions (v1.40.0). Fully reproducible report and writeup.

    Python

  3. obindoc obindoc Public

    Grounded RAG over a PDF with span-level citations, a verifier loop, and a tamper-evident HMAC-chained audit log. Under 900 lines, 29 tests.

    Python

  4. super-mcp-eval super-mcp-eval Public

    Evaluation harness for MCP servers and the agents that use them: schema compliance, tool-selection accuracy with Wilson 95% confidence intervals, DuckDB persistence, Streamlit dashboard.

    Python

  5. streamsense streamsense Public

    Multimodal human activity recognition on PAMAP2: a late-fusion CNN plus transformer with ONNX export, int8 quantization, and a FastAPI inference server for near-real-time inference on wearable IMU …

    Python

  6. agent-rx agent-rx Public

    Close the loop on agent-triage: diagnose, propose, A/B-test, and accept fixes for multi-agent failures, with a learned prioritizer that decides what's worth fixing.

    Python