Mohanish Kulkarni

I built AI-powered security and machine learning systems at Azpirantz Technologies LLP (InfosecTrain) and SAP Labs India. My work spans LLM-driven vulnerability scanning, RAG pipelines, ML prediction models, and real-time IoT data workflows.

At Azpirantz, I designed a web vulnerability scanner that classifies CVE and OWASP findings with ML, feeds them into a RAG pipeline, and uses an LLM agent to generate contextual remediation reports. I Dockerised the system with GitHub Actions CI/CD, and its architecture documentation became the R&D division's AI security reference.

At SAP Labs India, I built end-to-end ML pipelines across 10,000+ PHQ-9 and GAD-7 data points, where a Random Forest and XGBoost ensemble reduced RMSE by 17%. I also integrated Arduino and Raspberry Pi sensor streams and used ABAP scripting to cut report generation time by 30%.

I build practical agentic AI products, from a local-LLM penetration-testing platform to an MCP middleware package and an AI legal advisor. I enjoy taking ownership from ingestion and fine-tuning through retrieval, deployment, and frontend delivery.