Pratyush Kumar Singh

I’m an AI / Machine Learning engineer with a strong software-engineering foundation (~2 years backend at Nokia — Java & Python). I design and deploy production AI systems in Python, from data preprocessing through evaluation, validation, and cloud deployment.

I’ve built multi-model LLM pipelines with retrieval-augmented generation (RAG) and LLM integration using Gemini / Claude / OpenAI. My work focuses on reliability: I implement model validation & calibration loops and operational monitoring so AI keeps performing in real conditions.

Across my experience, I’ve applied statistical ML for prediction, anomaly detection, and clustering on large datasets. For example, I engineered a full analytics layer using Bayesian inference, CUSUM anomaly detection, gradient-boosted prediction, and EM / Gaussian-mixture clustering, persisting results across many database tables.

I’m most energized by turning complex requirements into systems that run 24/7—like my multi-LLM, 5-stage pipeline with model routing and a 110K-line, 180+ module architecture. I’m eager to apply machine learning and data analytics to solve complex, real-world problems.