Skip to content
View AshC2004's full-sized avatar
1 follower · 0 following

Block or report AshC2004

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
AshC2004/README.md

Ashish Chhabra

AI/ML Engineer | Building retrieval systems, agentic AI tools, and data-driven platforms.

Final-year AI & Data Science student at NMIT Bangalore. I build AI systems that work beyond the notebook -- RAG pipelines that don't hallucinate, tools that AI agents can reason with, and models that drive real decisions. I care about clean architecture, evaluation rigor, and ML that ships to production.

What I Build

AI/ML Systems

Project Description
DarkStore Analytics Platform AI-powered microservices platform for dark-store operations: demand forecasting (ARIMA/SARIMA/Holt-Winters/RF/GB), picker-route optimization, anomaly detection, product-affinity analysis, an LLM chatbot, and an MCP server. 8 FastAPI services behind a gateway, React dashboard, Docker Compose.
LLM Eval Harness Automated evaluation pipeline for RAG/agent outputs: groundedness scoring (n-gram overlap, TF-IDF cosine similarity), per-sentence hallucination detection, and cross-run regression tracking with a CI-friendly CLI. 63 tests, 87% coverage, no network calls required.
LLM Gateway Multi-provider LLM request gateway with routing (explicit, round-robin, fallback), exponential-backoff retries, per-request cost tracking, and SQLite request logging. FastAPI, dependency-injected providers, 27 tests.
Semantic SQL MCP Three-microservice MCP stack exposing a Postgres warehouse to Claude/Cursor: semantic column mapping with automatic JOIN-path discovery, Row-Level Security per analyst, AST-based SQL allowlisting. 13 MCP tools.
Conversation API Production-grade REST API for AI-powered conversations with FastAPI, Supabase, and multi-provider LLM support (Groq + Google AI). SSE streaming, token counting, cost tracking, ownership enforcement. 26 integration tests, full OpenAPI docs.
MCP Financial Server Model Context Protocol server exposing financial data tools over Supabase PostgreSQL. Query company profiles, financial reports, stock prices, and analyst ratings for Claude Desktop and Cursor.
Document Intelligence RAG pipeline using LangChain + Pinecone for technical Q&A. Cross-encoder re-ranking, source attribution, 65% hallucination reduction, sub-2s latency.
MCP Model Context Protocol server for data contract impact analysis. AI agents autonomously analyze schema changes and trace downstream dependencies.
Job Tracker MCP Server MCP server for tracking job applications through the hiring pipeline: status transitions, follow-up reminders, and pipeline summaries over SQLite.
Expense Tracker MCP Server MCP server for tracking personal expenses and wallet balance over SQLite. Input-validated tools, category taxonomy exposed as an MCP resource, full test suite.
Reddit MCP Server MCP server for browsing Reddit content in Claude: subreddit feeds, post content, and full comment-tree traversal, no API keys required.
TOON Token-efficient code notation for LLMs. AST-based compression that reduces token usage by 50-70%. Zero dependencies.

Data Science & Analytics

Project Description
NSE Stock Portfolio Analytics Predictive models for portfolio risk assessment (VaR, CVaR, Sharpe ratio) across 50+ NSE stocks. Power BI dashboards with 15+ KPIs.

Software Engineering

Project Description
Email Service Transactional email API with async processing (RabbitMQ), rate limiting (Redis), retry logic, and dead letter queues. Fully containerized.
How to GitHub Open-source guide to Git workflows -- 10 chapters from beginner to CI/CD.

Tech Stack

Languages        Python · JavaScript · SQL

AI/ML            LangChain · RAG · PyTorch · scikit-learn
                 Pinecone · OpenAI API · Groq · Google AI
                 MCP Protocol

Backend          FastAPI · Supabase · Docker · RabbitMQ
                 Redis · PostgreSQL · MongoDB · REST APIs
                 Server-Sent Events (SSE)

Data             Pandas · NumPy · Apache Kafka · Power BI
                 Statistical Modeling · ETL Pipelines

Currently

  • 🎓 B.E. AI & Data Science @ NMIT Bangalore, 8.5 CGPA, Graduating June 2026
  • 📜 Certified: Deep Learning Specialization (Coursera) · Bloomberg Technology Program
  • 🏛 VP @ AUGMENTiX (AI & Data Science society) · Secretary @ IEEE CIS

Open To

AI/ML Engineering: RAG, LLM integration, agentic AI, production ML pipelines Data Science: Predictive modeling, analytics, risk modeling, dashboards Software Engineering: Backend systems, APIs, distributed architecture

📧 ashchh2004@gmail.com · LinkedIn

Pinned Loading

  1. Email_Service Email_Service Public

    An Email Delivery Service : A transactional email API built with FastAPI, RabbitMQ, and PostgreSQL. This project demonstrates async message processing, reliable delivery, rate limiting, and event-d…

    Python

  2. TOON-Token-Object-Oriented-Notation- TOON-Token-Object-Oriented-Notation- Public

    Token-efficient notation for LLM context: AST-based compression that cuts token usage 50-70% with zero dependencies.

    Python

  3. conversation-api conversation-api Public

    A production-grade REST API for managing AI-powered conversations with streaming support, built with FastAPI and Supabase.

    Python

  4. document-intelligence document-intelligence Public

    RAG Project with document intelligence capabilities

    Python

  5. Dark-Store-Analytics-Platform Dark-Store-Analytics-Platform Public

    Analytics platform for dark-store/quick-commerce operations: demand forecasting, inventory optimization, and delivery performance dashboards.

    Python

  6. llm-eval-harness llm-eval-harness Public

    Automated groundedness scoring, hallucination detection, and regression tracking for RAG and agent outputs

    Python