Skip to content
View prabhaingit's full-sized avatar
0 followers · 1 following

Block or report prabhaingit

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
prabhaingit/README.md

Hi, I'm Prabhakaran Kuppusamy

Enterprise AI & Innovation Leader | Generative AI | Agentic AI | AI Engineering | AI Research

I build AI-native solutions for the enterprise, working where enterprise systems (especially SAP) meet generative AI, agents, retrieval and tabular foundation models.

I'm interested in turning complex enterprise data and processes into intelligent, explainable, AI-assisted systems, and in measuring honestly whether they work.

Selected projects

Precedent: explainable tabular AI

Row-level, removal-verified explanations for TabPFN predictions, served over MCP with an agent layer and a tamper-evident audit ledger. Tabular foundation models, explainability, MCP, agents and auditability in one project. prabhaingit/precedent

TabFM-Benchmark: tabular foundation model evaluation

Contributor to an open benchmark of tabular foundation models against tuned gradient-boosted trees and an MLP. I added the SAP-RPT-1 model (merged upstream), fixed a nested-parallelism slowdown in the LightGBM/XGBoost wrappers, and ran the 15-dataset, 5-seed evaluation. prabhaingit/TabFM-Benchmark (fork of Siri1702/TabFM-Benchmark)

SAP HANA VectorDB RAG

Retrieval-Augmented Generation on SAP HANA Cloud vector capabilities: document upload and embedding, similarity search and LLM-based answers, built with LangChain. prabhaingit/SAP-HANA-VectorDB-RAG

Domain-specific LLM for SAP HANA

Fine-tuning a StarCoder model on SAP HANA-specific SQL queries. prabhaingit/LLM-finetuned-HANA

Focus areas

  • Generative AI for enterprise applications
  • Agentic AI and MCP
  • Retrieval-Augmented Generation and vector search
  • Domain-specific LLMs
  • Explainable AI and auditability
  • AI evaluation and benchmarking
  • SAP + AI

Technology

Area Used in my projects
Languages & notebooks Python, Jupyter
Tabular ML TabPFN, SAP-RPT-1, XGBoost, LightGBM, CatBoost
GenAI LangChain, OpenAI APIs, StarCoder fine-tuning
Agents Model Context Protocol (MCP)
Enterprise SAP HANA Cloud (vector engine), SAP BTP

What I'm exploring now

How can AI systems move beyond answering questions to reasoning over enterprise knowledge, taking meaningful actions, and explaining and auditing what they did?

Enterprise data → knowledge → models → agents → tools (MCP) → action, with explainability and governance throughout.

Let's connect

Open to conversations on enterprise generative AI, agentic AI, RAG, AI evaluation, SAP + AI and explainable AI.

Pinned Loading

  1. SAP-HANA-VectorDB-RAG SAP-HANA-VectorDB-RAG Public

    Retrieval-Augmented Generation using SAP HANA Cloud vector capabilities for enterprise knowledge applications.

    Python 2

  2. LLM-finetuned-HANA LLM-finetuned-HANA Public

    Experimenting with domain-specific LLM fine-tuning (StarCoder) for SAP HANA SQL and enterprise code generation.

    Jupyter Notebook

  3. precedent precedent Public

    Explainable TabPFN predictions using verified row-level precedents, MCP, agents and tamper-evident auditability.

    Jupyter Notebook

  4. TabFM-Benchmark TabFM-Benchmark Public

    Forked from Siri1702/TabFM-Benchmark

    Fork of Siri1702/TabFM-Benchmark. Contributed SAP-RPT-1 integration (merged upstream), benchmark fixes and a 15-dataset tabular foundation model evaluation.

    Jupyter Notebook