Looking for a job

Harsh Giri

@harshgiri

Message

Machine learning engineer building agentic RAG systems and clinical prediction models with Microsoft SAP Edunet.

India

Message

What I'm looking for

Seeking a remote/hybrid AI/ML internship where I can apply my experience with deep learning, NLP, and RAG pipeline development on real-world problems. I'm looking to work with production-grade ML systems and agentic AI workflows, ideally on a small team where I can contribute directly and learn from senior engineers. Open to 3-6 month internships alongside my studies

At Microsoft SAP Edunet, I engineered a multi-disease clinical prediction system using 45,000+ patient records. By combining Random Forest, SVM, and neural networks, I improved accuracy from 82% to 93.7% and increased high-risk minority-class recall by 23%.

For my Enterprise Grade Agentic RAG project, I built a guardrailed Q&A service with FastAPI and LangGraph, combining two-stage Qdrant retrieval with NeMo Guardrails. I also contributed production bug fixes and CI/CD regression testing workflows to Apache OpenOffice.

Experience

Work history, roles, and key accomplishments

Open-Source Contributor

Apache OpenOffice / GirlScript Summer of Code

Jan 2024 - Jan 2026 (2 years)

Contributed 5+ production bug fixes across C++, Java, and XML in Apache OpenOffice, leveraging AST-based static code analysis across 15+ modules. Implemented CI/CD regression testing workflows integrated with the Apache build system, reducing post-merge defect rates by 25%.

C++Java XML Static Code Analysis CI CD Regression Testing Doxygen

ML Apprentice

Microsoft SAP Edunet

Jan 2025 - Mar 2025 (2 months)

Engineered a multi-disease clinical prediction system on 45,000+ patient records, ensembling Random Forest, SVM, and Neural Networks to boost accuracy from 82% to 93.7%. Automated feature extraction via PCA and 10-fold cross-validation, implementing SMOTE to improve high-risk minority-class recall by 23%.

Random Forest SVM Neural Networks PCA SMOTE Cross Validation Feature Engineering Clinical Risk Prediction