Skip to content
View Shashankss1205's full-sized avatar
🎯
Focusing
🎯
Focusing

Organizations

@gc-os-ai

Block or report Shashankss1205

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.

Content in all repositories owned by your account will be closed.
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
Shashankss1205/README.md

Hey, I'm Shashank 👋

I teach machines to read ancient manuscripts, understand 330+ languages, understand spaghetti code, and stop hallucinating about Bitcoin.

🔬 Research Scientist @ American Express  •  🛠️ Maintainer of CodeGraphContext  •  🎓 IIT (BHU) Varanasi

profile views


🧠 What I actually do

I build things at the intersection of LLMs, agents, and retrieval — and I like my AI systems the way I like my code: grounded, fast, and reproducible.

class Shashank:
    def __init__(self):
        self.role = "Research Scientist @ AmEx"
        self.obsessions = ["Agentic AI", "RAG", "MCP", "Inference optimization"]
        self.currently = "Making agentic evaluations better and automated"
        self.side_quest = "Mentoring GSoC 2026 contributors @ HumanAI"

    def fun_fact(self):
        return "My OCR models read 200-year-old Spanish documents better than I read my own handwriting"

⭐ The project I'm proudest of

CodeGraphContext — a CLI toolkit + MCP server that indexes your codebase into a knowledge graph so AI assistants can actually understand your code instead of guessing.

📦 225k+ downloads · ⭐ 4k+ stars · 👥 150+ contributors — all in 12 months

🗺️ The journey so far

When What Where
2026 → Research Scientist — Agent Evals American Express
2026 Built sktime-mcp — an MCP engine for time-series forecasting Sktime
2025 & 2024 GSoC Contributor → now Mentor — OCR for historical archives (97% accuracy) HumanAI
2025 RAG for Bitcoin dev tooling · SQLite engine in C++ Summer of Bitcoin
2025 Multilingual search across 53k storybooks in 330+ languages Pratham Books (C4GT)
2024 PII anonymization pipelines for telecom data Linux Foundation (LFX)
2024 Fashion trend prediction with transfer learning Blue Yonder

🏆 Occasionally I win things

🥇 NLP.py @ IIT Kharagpur · 🥇 National Winner, Blue Yonder Crystal Ball · 🏅 Best Knowledge Graph RAG @ Hypermode Challenge · 🥈 Low Code ML @ IIM Ahmedabad · 📄 Published: OCR Approaches for Humanities (95% char-level accuracy) · 📊 Kaggle dataset featured on Global Trending

🧰 Weapons of choice

Core: Python · C/C++ · TypeScript · SQL

AI/ML: PyTorch · JAX · Hugging Face · vLLM · Unsloth

Agents & RAG: LangGraph · LlamaIndex · DSPy · MCP SDK · ChromaDB · FAISS · Neo4j

Infra: FastAPI · Docker · PostgreSQL · GCP

📊 Numbers, because ML people love numbers

streak

📫 Let's talk

⚡ Ask me about RAG, MCP servers, or why your agent keeps calling the wrong tool.

Pinned Loading

  1. CodeGraphContext/CodeGraphContext CodeGraphContext/CodeGraphContext Public

    An MCP server plus a CLI tool that indexes local code into a graph database to provide context to AI assistants.

    Python 4.2k 829