About Me

Hi, I’m Soumya - I work on AI systems, with a focus on computer vision. I enjoy building models, experimenting with research ideas, and writing bugs so creative they feel intentional. Most of my workflow involves training networks, reading papers, and pretending final_final_v3 is a perfectly reasonable folder name.

Education

  1. Maulana Abul Kalam Azad University Of Technology

    2023 — 2027

    Bachelor of Technology - BTech, Computer Science Engineering & Business Systems (CSE&BS)

  2. Indian Institute Of Technology Madras

    2023 — 2027

    Bachelor of Science - BS, Data Science & Applications

Skills

  • Python
  • LaTeX
  • Flask
  • C
  • C++
  • C#
  • HTML5
  • CSS3
  • JavaScript
  • Bootstrap
  • MS Office
  • G Suite
  • Visual Studio
  • MATLAB
  • NumPy

Experience

Academic & Research

  1. Research Trainee — Variable Energy Cyclotron Centre (VECC), Government of India (GoI)

    Mar 2025 — Oct 2025

    Built foundations in ML/DL; literature survey and practical implementations for Talking Head Generation.

Industry & Entrepreneurship

  1. Co-Founder — MSG Group

    Apr 2019 — Present

    Co-founded to help underprivileged sellers reach international markets; expanded into software development and publications.

  2. Co-Founder — Oniroop Herbal

    Jun 2022 — Present

    Led strategic partnerships and product development; integrated sustainable practices in herbal wellness.

  3. Market Research Analyst — JITSIE, IIT Madras

    Sep 2024 — Mar 2025

    Supported investment narratives with market intelligence for deep-tech ideas and ventures.

  4. Head — Technical Department, Adhyayan (LearnAdhyayan)

    Aug 2021 — Aug 2023

    Built and deployed customer-facing systems; reduced downtime by ~50% in 6 months; developed a new website.

Other Activities

  1. Eastern Zonal Secretary & Technical Recruiter — Dhyan Chand Sports Council, IIT Madras

    Jul 2023 — Dec 2023

    Coordinated Savaan 2023; promoted inclusive participation across Eastern Zone; interviewer for 100+ candidates.

  2. Public Relations — IITM Diplomacia

    Jul 2023 — Sep 2023

    Drove PR campaigns with 111% engagement growth and 18% higher brand recognition in one month.

Publications

2026

  • Vineet Kumar Rakesh, Soumya Mazumdar, Tapas Samanta, Himanshu Kumar Pandey, Amitabha Das, & Sarbajit Pal, “Analysis of Hyperparameter Optimization Effects on Lightweight Deep Models for Real-Time Image Classification,” Scientific Reports, 2026, doi:10.1038/s41598-026-42748-w
  • Vineet Kumar Rakesh, Soumya Mazumdar, Research Pratim Maity, Sarbajit Pal, Amitabha Das, & Tapas Samanta, “Advancing Talking Head Generation: A Comprehensive Survey of Multi-Modal Methodologies, Datasets, Evaluation Metrics, and Loss Functions,” The Visual Computer, vol. 42, no. 9, 2026, doi:10.1007/s00371-025-04232-w
  • Soumya Mazumdar, Vineet Kumar Rakesh, & Tapas Samanta, “PrivFedTalk: Privacy-Aware Federated Diffusion with Identity-Stable Adapters for Personalized Talking-Head Generation,” IEEE Transactions on Circuits and Systems for Video Technology, under review, 2026. arXiv:2604.08037
  • Vineet Kumar Rakesh, Soumya Mazumdar, Tapas Samanta, Himanshu Kumar Pandey, Amitabha Das, & Sarbajit Pal, “Adaptive Video Conferencing Under Severe Bandwidth Constraints Using Audio-Driven Talking-Head Reconstruction,” Engineering, under review, 2026. arXiv:2602.12758
  • Vineet Kumar Rakesh, Abhijit Bhattacharjee, Soumya Mazumdar, Tapas Samanta, Himanshu Kumar Pandey, Amitabha Das, & Sarbajit Pal, “Symbolic Vedic Computation for Low-Resource Talking-Head Generation in Educational Avatars,” IEEE Transactions on Multimedia, under review, 2026. arXiv:2602.08775
  • Vineet Kumar Rakesh, Soumya Mazumdar, Abhijit Bhattacharjee, Tapas Samanta, Himanshu Kumar Pandey, Amitabha Das, & Sarbajit Pal, “Bandwidth-Aware Performance Analysis of Video Conferencing Platforms,” Multimedia Tools and Applications, under review, 2026.
  • Soumya Mazumdar, S. Saha, M. Roy, S. Kundu, P. K. Halder, S. Das, U. Adhikari, & L. Nayak, “Handheld Device for Real-Time Neural Network Analysis,” Patent No. 490413–001, Controller General of Patents, Designs & Trade Marks, Kolkata, 2026. Official source
  • Soumya Mazumdar, Vineet Kumar Rakesh, & Tapas Samanta, “BayesFusion-SDF: Probabilistic Signed Distance Fusion with View Planning on CPU,” IEEE 5th International Conference on Intelligent Reality (ICIR 2026), accepted, 2026. arXiv:2602.19697
  • Vineet Kumar Rakesh, Soumya Mazumdar, Research Pratim Maity, Tapas Samanta, Himanshu Kumar Pandey, & Amitabha Das, “Empirical Evaluation of State-of-the-Art Talking Head Generation Models,” in International Conference on Recent Advances in Artificial Intelligence & Smart Applications (RAAISA 2025), Lecture Notes in Networks and Systems, Springer, 2026. Yet to publish.
  • Vineet Kumar Rakesh, Mo Moin, B. Sarkar, Soumya Mazumdar, Tapas Samanta, Himanshu Kumar Pandey, & Amitabha Das, “THGNLR: An Intelligent Application for Automated, Large-Scale Multimodal Human Audio-Video Data Acquisition and Auto-Cleansing,” in International Conference on Advancement of Intelligent Computing in Data Analysis and Knowledge Discovery (AICDAKD 2025), Lecture Notes in Networks and Systems, Springer, 2026. Yet to publish.
  • Soumya Mazumdar & Vineet Kumar Rakesh, “TempoSyncDiff: Distilled Temporally-Consistent Diffusion for Low-Latency Audio-Driven Talking Head Generation,” 19th European Conference on Computer Vision (ECCV 2026), under review, 2026. arXiv:2603.06057
  • Soumya Mazumdar & H. Hosni, “Machine Learning and Deep Learning: Necessary Techniques for Extracting Meaningful Information in Large Databases,” in Next Generation Blockchain for Next Generation Society with Futuristic Technologies, CRC Press (Taylor & Francis), 2026, doi:10.1201/9781003620822-20
  • Vineet Kumar Rakesh, Soumya Mazumdar, Research Pratim Maity, Sarbajit Pal, Amitabha Das, & Tapas Samanta, “Transparent and Trustworthy Data Practices for Talking Head Generation: A Comprehensive Survey of Datasets and Ethical Implications,” in Centering Transparency and Trust in Data and AI Ecosystems, IGI Global Scientific Publishing, 2026. Yet to publish.

2025

  • Vineet Kumar Rakesh, Soumya Mazumdar, Research Pratim Maity, Sarbajit Pal, Amitabha Das, & Tapas Samanta, “Quantitative Assessment in Talking Head Generation: A Comprehensive Review of Metrics and Loss Functions,” IET Software, under review, 2025.
  • Soumya Mazumdar, “Educational Implications of 6G Technology for Society 5.0,” in Human-Centric Integration of 6G-Enabled Technologies for Modern Society, Elsevier, 2025, pp. 255–265, doi:10.1016/B978-0-443-27434-3.00017-9
  • Soumya Mazumdar & Itu Chowdhury, “Integration of Machine Learning, Deep Learning, and Internet of Things Technologies for Stress Management and Identification: A Systematic Analysis,” in Human-Centric Integration of Next-Generation Data Science and Blockchain Technology, Elsevier, 2025, pp. 407–418, doi:10.1016/B978-0-443-33498-6.00023-6
  • H. Hosni & Soumya Mazumdar, “6G-Enabled Robotics: The Future of Connectivity in Smart Industrial Manufacturing,” in Industrial Robotics in Smart Manufacturing, CRC Press (Taylor & Francis), 2025, pp. 241–258, doi:10.1201/9781003614470-12

Contact

Contact Form