R Ajay Kumar

I'm a Computer Science undergraduate and current Machine Learning Intern at Likwid AI, a B2B ERP platform for inventory and procurement management. I built and productionized an inventory demand and purchase-order prediction engine that hit ~87% accuracy and replaced roughly 60% of the company's manual, spreadsheet-based purchasing decisions — their first production ML system, built from scratch. I designed the FastAPI inference service to be reusable, so the same scaffolding now powers a second independent prediction module, and I redesigned the underlying data schemas and query patterns to eliminate latency bottlenecks between ingestion and serving. I'm currently architecting a RAG pipeline for autonomous ERP customization.

Outside work, I take my ML projects all the way to production rather than stopping at a notebook. I built a multilingual (English, Hindi, Tamil, Telugu) voice-and-text RAG chatbot with hybrid BM25 + semantic search, deployed on Azure Container Apps with CI/CD and a cross-platform React Native frontend. I also built a paper-trading equity system with a Black-Scholes options-exposure engine in C++ (via PyBind11), Monte Carlo stress testing over 10K paths, and 56 automated tests, plus an F1 race-outcome predictor (XGBoost) that hit ~85% winner-classification accuracy validated live during real race weekends.

I'm looking for full-time Junior ML/AI Engineer roles — remote preferred — where I can keep owning ML work end-to-end: modeling, inference APIs, and production deployment on Azure/GCP.