Skip to content

Open to new opportunities

Mukul Kumar - Senior Data Scientist

I build and ship deep learning systems for computer vision and NLP - from research prototype to production.

View projects Get in touch Download résumé

About

Hi, I'm Mukul.

I'm a Machine Learning and AI Engineer with 8 years of experience taking deep learning systems from research to production - most recently automating vehicle damage assessment for insurance claims, and before that, extracting structured data from millions of scanned receipts deployed across multiple countries.

My research interests are Graph Neural Networks and attention methods. That work has produced 5 filed patents and 2 peer-reviewed papers, including a lightweight GAT-based text line detector presented at the KDD Document Intelligence Workshop.

I care about the unglamorous parts that make models useful: clean data pipelines, honest evaluation, retraining when data drifts, and inference that fits the latency and cost budget (TensorRT, Triton, DeepStream). I'm always happy to collaborate with researchers on NLP or computer vision projects.

Role
Senior Data Scientist
Focus
Computer Vision, NLP, GNNs
Location
Bengaluru, India
Status
Open to new opportunities

  • 8

    Years in applied ML

  • 5

    Patents filed

  • 2

    Peer-reviewed papers

  • 45%

    Inspection cost reduction

Career

Experience

  1. Sep 2023 - May 2026 · Remote

    Senior Data Scientist · XA Group (Xpress Automation)

    • Led full-stack development of deep learning systems for automotive insurance claims, cutting manual inspection costs by 45%.
    • Designed ensemble CV models for damage and missing-part detection with 90% damage-detection accuracy.
    • Built self-supervised learning pipelines that reduced annotation costs by 30% annually; maintained production models and retrained on data drift.
    • PyTorch
    • AzureML
    • MLflow
    • Hugging Face
    • Ultralytics
    • TensorRT
    • Docker

  2. Apr 2021 - Aug 2023 · Chennai / Remote

    Senior Data Scientist · Blackstraw.ai

    • Consulted with client technical teams on data science and ML problems; lead developer on receipt information extraction deployed in multiple countries.
    • Combined CV models with NLP (NER, coreference, dependency parsing, knowledge-graph embeddings) to structure extracted data - 4 patents and a KDD workshop paper.
    • Technical interview panelist for data science hiring.
    • PyTorch
    • TensorFlow
    • DGL
    • C++
    • PyCUDA
    • TF Serving
    • TensorRT

  3. Nov 2019 - Mar 2021 · Chennai, India

    Data Scientist · Blackstraw.ai

    • Built a real-time social-distancing monitoring system across 40+ CCTV cameras with NVIDIA DeepStream and TensorRT; patent filed in India.
    • Developed computer-vision models for the receipt extraction pipeline.
    • DeepStream
    • TensorRT
    • GStreamer
    • FFmpeg
    • Docker

  4. Apr 2019 - Oct 2019 · Chennai, India

    Junior Data Scientist · Blackstraw.ai

    • Started on document-understanding and computer-vision projects for enterprise clients.

  5. Sep 2018 - Feb 2019 · Bengaluru, India

    Data Science Intern · CrowdANALYTIX

    • Designed and deployed deep learning models for dataX.ai, an AI-based fashion catalogue platform - from dataset preparation and GPU training to TensorFlow Serving.

  6. Jun 2017 - Jul 2017 · Dehradun, India

    Software Engineering Intern · Oil and Natural Gas Corporation (ONGC)

    • Built an Android employee-directory app with direct calling, bookmarks and Google sign-in.

Selected work

Projects

2023 – 2026Featured

Automated Vehicle Damage Assessment

Ensemble computer-vision system that detects vehicle damage and missing parts for insurance claims - 90% damage-detection accuracy.

  • Computer Vision
  • Self-Supervised
  • PyTorch
  • TensorRT
  • AzureML
Read case study
2019 – 2023Featured

Information Extraction from Receipts

CV + NLP pipeline that extracts and decodes purchased items from scanned receipts. Deployed in multiple countries; 4 patents and a KDD workshop paper.

  • Document AI
  • GNN
  • NLP
  • OCR
  • DGL
  • TensorRT
Read case study
2022Featured

GAT for Text Line Detection

4M-parameter Graph Attention + GRU model for line detection on receipts - state-of-the-art 97.23% F1 on CORD with 100–200× fewer parameters.

  • Research
  • Graph Attention Networks
  • Document AI
Read case study
2020

Social Distancing Monitoring System

Real-time, multi-camera system that flags social-distancing violations from existing CCTV feeds.

  • Computer Vision
  • DeepStream
  • TensorRT
  • Video Analytics
Read case study
2023

Parkinson's Detection from Voice

Vocal Tab Transformer that classifies Parkinson's disease from dysphonia features, beating gradient-boosted trees by 1%+ AUC.

  • Research
  • Transformers
  • Healthcare
  • Tabular
Read case study
2018 – 2019

AI Catalogue Models for dataX.ai

Deep learning models powering dataX.ai, a web platform that auto-generates fashion product catalogues.

  • Computer Vision
  • TensorFlow
  • TF Serving
Read case study

Research

Publications & patents

Peer-reviewed papers

Patents filed

  1. Methods, Systems, Apparatus and Articles of Manufacture to Detect Lines on Documents

    Filed Jun 2022 · US81259087

  2. Automatic Document Content Extraction and Decoding

    Filed Jun 2021 · US81254014

  3. Automated Extraction of Purchased Items in Images of Receipts Using Computer Vision and Deep Learning

    Filed Jun 2021

  4. Automated Receipt Decoding: Product Matching and Dictionaries

    Filed Jun 2021

  5. Method and System to Detect Distance Between Entities

    Filed Aug 2020 · IN 202021034381

Toolkit

What I work with

Models & Methods

  • Transformers & GPT
  • Diffusion & GANs
  • CNNs & GNNs
  • LSTMs
  • Self- & weakly-supervised learning

Frameworks

  • PyTorch
  • TensorFlow / Keras
  • Hugging Face
  • Deep Graph Library
  • OpenCV
  • ONNX
  • scikit-learn
  • spaCy

Deployment & MLOps

  • TensorRT
  • Triton Inference Server
  • TF Serving
  • NVIDIA DeepStream
  • Docker
  • AzureML / MLflow

Engineering

  • Python
  • C / C++
  • PyCUDA
  • SQL / MySQL
  • Bash
  • AWS / GCP

Background

Education & certifications

  1. 2024 - 2026

    M.Tech, Artificial Intelligence & Machine Learning

    Birla Institute of Technology and Science (BITS), Pilani - WILP

  2. 2014 - 2018 · Roorkee, India

    B.Tech, Computer Science & Engineering

    College of Engineering Roorkee

  • Qualified GATE CSE 2018 (score 346)
  • DeepLearning.AI TensorFlow Developer (2020)
  • Google Cloud Platform Fundamentals: Core Infrastructure
  • Observability Fundamentals - Elastic Stack
  • Data Science: Deep Learning in Python

Contact

Let's build something together

Have a role, a research collaboration or an interesting CV/NLP problem in mind? Send me a message and I'll reply as soon as I can.

Or email me directly at mukul.kr99@gmail.com.