Deepak Pradhan

Experience

1. Machine Learning Engineer

   Frinks AI

   Nov 2021 — Present · Bengaluru, IN

   Developed a real-time tracking and counting system for cement bags using YOLOv8 and OCSORT, achieving high accuracy in tracking and counting cement bags on conveyor belts, processing 300k bags daily. • Implemented a 3D dimensioning solution using Open3D and YOLOv8 to measure object dimensions at the micron level from point cloud data. • Designed a quality inspection model using PointRend to analyze sheets, achieving 100% defect detection accuracy on 10k+ samples, and reducing manual QA time by 60%. • Deployed an automatic number plate recognition (ANPR) system using YOLOv8 and PaddleOCR, trained on 15k+ custom images, achieving 100% accuracy across varied lighting and plate conditions, and recognizing 200+ vehicles daily across multiple gates. • Implemented a robust anomaly detection system for industrial inspection using U-Net++, Anomalib with PatchCore, and EfficientAD, achieving 100% precision in detecting rare and unknown defects across 50k+ inspection samples. • Developing a custom question-answering chatbot using LLama3.2 and Langchain, fine-tuned with PEFT LoRA to enable efficient domain-specific retrieval-based answering.

Education

• upGrad.com

  PGDDS, DATA SCIENCE

  2019 — 2020

• NICMAR University

  Master’s Degree, Advanced Construction Management

  2015 — 2017

• International Institute of Information Technology Bangalore

  PGDDS, Data Science

  2019 — 2020

• International Institute of Information Technology Bangalore

  PGDDS, Data Science

• Kalinga Institute of Industrial Technology, Bhubaneswar

  Bachelor’s Degree, Civil Engineering

  2010 — 2014