👋 Hello! I’m Anubhav, a Manager at HSBC specializing in Generative AI, RAG and production-grade ML systems.

[

1. About Me
2. Highlights
3. Skills
4. Featured Projects
5. Try It Yourself
6. How To Contribute
7. Contact

• Experience with ETL pipelines, RAG, LangChain, LangGraph, AutoGen, and agentic AI.
• Formerly at Capgemini delivering scalable enterprise data workflows.

"I solve hard data problems with intelligent pipelines and responsible AI."

• Built agentic workflows for multi-step reasoning over knowledge sources (Wikipedia, ArXiv, Tavily).
• Developed Swarm AI orchestration using n8n + LLM agents with real-world integrations (Gmail, Calendar, Travel booking).
• Engineered low-latency RAG pipelines with Groq LPU + Llama/Gemma, backed by Pinecone/AstraDB.
• Deployed APIs using FastAPI + LangServe, monitored with LangSmith.

Python: Proficient in Python for data science, machine learning, and statistical analysis.
SQL: Extensive experience working with relational databases, creating and optimizing queries.
Pandas: Expertise in data manipulation and analysis.
NumPy: Advanced numerical computing and handling of large datasets.
Scikit-Learn: Building machine learning models, including regression, classification, and clustering algorithms.
Matplotlib & Seaborn: Data visualization for insightful analysis and clear communication.
Power BI: Creating interactive dashboards and reports for business intelligence.

• Postman, Gen AI, LLM, NLP, Streamlit, Computer Vision
• Agentic AI, LangGraph, Prompt Engineering, LLM Fine-Tuning
• Git, Excel, Jira

• Designed an agentic AI workflow using LangGraph StateGraph, enabling structured orchestration of LLM reasoning.
• Implemented tool-enabled agents with LangChain, integrating external sources (Wikipedia, ArXiv, Tavily) for dynamic retrieval.
• Built conditional routing and multi-step reasoning pipelines so LLMs autonomously decide when to invoke tools.

• Designed an event-driven multi-agent architecture in n8n where AI agents coordinate for workflows (email → calendar → travel booking).
• Developed modular sub-workflows with API integrations (Google Workspace, search APIs) and dynamic prompt chaining.
• Deployed a real-world assistant with Telegram + voice interface, showing end-to-end automation and multimodal orchestration.

• Built high-performance RAG pipeline with Groq LPU using Llama 3.1 / Gemma for fast query responses over large unstructured content.
• Designed scalable vector search with AstraDB (Cassandra) and Pinecone, optimizing embeddings for context-aware generation.
• Developed end-to-end pipelines and deployed LLM apps as REST APIs using FastAPI + LangServe, with monitoring via LangSmith.

1. Clone:

git clone https://github.com/your-username/your-repo.git
cd your-repo

2. Setup venv:

python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt

3. Run a sample pipeline:

python examples/run_rag_demo.py

4. Explore project notebooks in notebooks/ for interactive experimentation.

• Star ⭐ the repository if you find it useful.
• Open issues for bugs or enhancement ideas.
• Submit PRs with clear descriptions and tests.

• Email: shanukumar.srivastava77@gmail.com

Enjoy exploring — let’s make AI & ML in finance smarter and safer together!