Utkarsh — About

Hi, I’m Utkarsh.

Data Scientist & AI/ML Engineer · Computer Vision · NLP · Generative AI · MLOps

Bio

Data Scientist and AI/ML Engineer with 3+ years building end-to-end ML systems across computer vision, NLP, generative AI, and big data — from social-impact and geospatial domains to multimodal search and LLM applications. Grounded in the full data engineering stack: Hadoop, Spark, Hive, cloud pipelines, and ETL at scale. Currently deepening work in transformer architectures, RAG systems, embedding fine-tuning, and LLM fine-tuning.

Comfortable moving between IC model work and leading cross-functional teams (up to 65 contributors). Treats MLOps, reproducibility, versioning, and pipeline hygiene as core engineering, not an afterthought. Away from the keyboard: cricket, hockey, MMA, the NFL, and movies.

Skills

ML / AI

Computer Vision YOLO SAM NLP Generative AI RAG Embedding Fine-tuning LoRA Matryoshka/MRL Transformers LLM Fine-tuning PyTorch TensorFlow

Big Data & Pipelines

Hadoop Spark Hive Airflow ETL/ELT Delta Lake

MLOps

DVC MLflow Git Docker experiment tracking

Data & Cloud

Python SQL Pandas Tableau AWS Azure Streamlit Gradio Hugging Face

Languages

Python SQL R Go

Experience

Data Scientist / Product Manager — Omdena

Omdena · Project-based May 2022 — Present

  • Architected and deployed end-to-end ML solutions across CV, NLP, and generative AI, handling data pipeline design through model deployment for multiple production use cases.
  • Led project delivery for multi-disciplinary teams (up to 65 contributors), coordinating engineers, domain experts, and business stakeholders to ship on time.
  • Established team-wide MLOps standards using Git, DVC, and MLflow, reducing experiment chaos and making model reproduction a non-issue.
  • Translated model performance into business-facing insights; owned stakeholder communication from progress updates to final deliverables.

Cricket Video Analyst — Mad About Sports

Mad About Sports · Intern 2022

  • Built player performance analytics in Python (Pandas, NumPy, Matplotlib, Seaborn) and Tableau from raw video and match data.
  • Developed predictive models for batting/bowling performance trends; findings fed directly into coaching and selection strategy.

Education

PG Certificate Programme in Big Data Analytics (PGCP-BDA)

C-DAC 2026

  • Big Data Engineering: Hadoop ecosystem (HDFS, MapReduce, YARN), Hive, HBase, Spark, Airflow-based ETL, Data Lakes & Lakehouses (Delta Lake).
  • Deep Learning & Generative AI: CNNs, RNNs, transfer learning, object detection/segmentation; Transformers, BERT, LLMs, fine-tuning & alignment (RLHF), SLM/LLM deployment.

B.Tech — Information Technology

GGSIPU, Delhi 2017 — 2021

GPA 7.7 · Relevant coursework: Data Structures, Algorithms, Database Systems, Statistics.

Off the clock

Cricket Hockey MMA NFL Movies

Get in touch

The fastest way to reach me is LinkedIn — or say hi on any platform below. Always happy to talk machine learning, RAG, embeddings, cricket analytics, or trade movie recommendations.

Back to top