Vedant Wedhane

I’m a versatile AI/ML engineer with broad coverage across NLP, computer vision, time-series, and generative AI. I focus on turning models into working systems—business-ready pipelines, reliable services, and measurable impact.

At Rajvil IT Services, I deployed a production NLP classification workflow that automated SAP IBP alert triage, reducing manual processing by 40%. I built end-to-end ML pipelines (data → model → FastAPI deployment) and translated stakeholder requirements into RCA tooling for supply chain analysis using Python and HuggingFace classification pipelines.

I’ve also delivered applied time-series ML for energy consumption forecasting at Infosys (Springboard), strengthening feature engineering, evaluation, and tuning with Scikit-learn pipelines. For real-world computer vision, I built an LMS proctoring system using OpenCV and CNNs for gaze tracking, object detection, and behavioral analysis—an effort that was funded and Springer-published (BIDA 2025).

On the GenAI side, I’ve built agentic RAG solutions with LangChain, emphasizing citation enforcement to drive near-zero hallucinations. I bring full ML stack proficiency across Python, PyTorch, TensorFlow, HuggingFace, FastAPI, Docker, and cloud tooling (AWS and certified GCP), and I’m continuously deepening my expertise through my IIT Roorkee AI Engineering on Cloud & AIOps PG certification.