prabhat shekhar
Machine Learning Engineer
About
AI/ML researcher and engineer with full-cycle, hands-on experience building, training, and statistically validating deep learning systems on complex, high-dimensionality, real-world data — including human-derived data across text, speech, and imaging. Delivered a novel two-stage inference pipeline for multi-document summarisation that outperformed six published state-of-the-art baselines, with results validated through rigorous ablation studies and blind human evaluation using inter-annotator agreement starsRcs. Comfortable owning the full ML lifecycle end-to-end: data pipeline design, model training and debugging at scale, statistical evaluation, and reproducible experimentation — with a track record of translating research prototypes into measurable, benchmarked outcomes. Seeking to apply this foundation in transformer architectures, generative AI, and applied deep learning to production-grade ML solutions for pharma and healthcare applications.
Skills
Want to reach prabhat?
Companies and referrers connect through Kaam Paao, where every introduction is trusted and tracked.