Background

Experience

  1. 2015 – present

    Samsung R&D Institute India – Bangalore (SRI-B)

    GenAI Research Scientist & Architect, Machine Learning — Visual GenAI Group, Visual Intelligence Team (2026 – present) · Advanced Technology Lab, Advance Research Group (2015 – 2025)

    1. Chief Engineer 2015 – 2020
    2. Senior Chief Engineer 2020 – 2022
    3. Architect (Staff Engineer grade) 2022 – present
    • Led the Advanced Technology Lab until 2025 — reporting manager for multiple teams and projects; since 2026, one of the lead GenAI research scientists and architects in the Visual GenAI Group.
    • Lead the team building the vision-language model behind natural-language search and other Gallery features; commercialised in the Galaxy S25 series with Samsung Research (Seoul) and the Mobile eXperience division (Suwon).
    • Lead development of multimodal LLMs and Large Action Models for agentic AI in mobile applications.
    • Led the on-device personalised knowledge graph from research to commercial release (Galaxy S21 series onward) with Samsung’s Big Data Center.
    • Earlier: multimodal Story/Memory creation in Gallery, natural expression generation for a social robot, on-device Smart Reply for Galaxy Watch, smart assistant for mobile.

  2. May 2014 – Apr 2015

    ERCIM Postdoctoral Research Fellow

    Institute of Informatics and Telematics (IIT‑CNR), National Research Council of Italy, Pisa · ERCIM Alain Bensoussan Fellowship Programme

    • Biomedical question answering: linking clinical cases to the literature relevant for patient care, with Dr. Marina Buzzi. Information retrieval, information extraction, NLP and knowledge graphs.
    • Published at ITBAM 2015 (Springer LNCS). Research exchange at Universitat Politècnica de València with Dr. Paolo Rosso (Feb 2015).

  3. 2007 – 2014

    Jadavpur University

    Dept. of Computer Science & Engineering, Kolkata

    • Senior Research Engineer Aug 2012 – Apr 2014

      Cross-Lingual Information Access (CLIA) Phase II, the Government of India’s “Sandhan” Indian-language search engine (Lucene, Nutch, Solr). Improved snippet and summary generation; built question answering and web-page summarisation; senior member of the project team (documentation, presentations, project management).

    • Junior Research Fellow Sep 2010 – Jul 2012

      Sole research fellow on the Indo-French (IFCPAR) project “An Advanced Platform for Question Answering Systems”, with Prof. Sivaji Bandyopadhyay (Jadavpur) and Dr. Patrick Saint-Dizier (IRIT-CNRS), including a research stay at IRIT, Toulouse (Sep – Oct 2011).

    • Junior Research Engineer Feb 2007 – Aug 2010

      CLIA Phase I. Built the snippet-generation and summary-generation modules of the search engine.

    • Guest Faculty Jul 2011 – Apr 2014

      Taught B.E. and M.Tech. courses (programming languages, numerical analysis, operating systems lab) and assisted Master’s theses.

What I do

  • Vision-language models for retrieval and understanding
  • Multimodal LLMs and agentic systems (Large Action Models)
  • On-device machine learning under mobile constraints
  • Knowledge graphs and personalisation
  • NLP: question answering, summarisation, retrieval, evaluation
  • Research-to-product leadership: building and managing research teams, cross-site delivery

Tools Python · TensorFlow · Java · SQL

Invited talks

  1. 2 July 2026

    The Future of Work

    Distinguished lecture — BIM 43 Inauguration & Orientation Programme, Bharathidasan Institute of Management (BIM), Tiruchirappalli

    A lecture to the incoming MBA batch on the future of work: the five forces reshaping it, generative AI first among them, and what the PwC, World Economic Forum and Goldman Sachs figures say about jobs, skills and wages. Then how AI products are actually built — the on-device knowledge graph and natural-language search in Gallery, which went from the lab to Galaxy phones, and the agentic AI now following them — and the T-shaped, AI-fluent manager all of this asks for. Followed by Q&A.

    • Dr. Pinaki Bhaskar receiving a memento from Prof. Asit K. Barma, Director of BIM Tiruchirappalli, in front of the lecture slides.
      With the Director, Prof. Asit K. Barma
    • Welcome banner for the inauguration and orientation programme of BIM Batch 43, with portraits of the Director and the invited guests, Dr. Pinaki Bhaskar among them.
      The programme banner (orientation from 25 June)
    • Dr. Pinaki Bhaskar with the Director, Prof. Asit K. Barma, and two colleagues from BIM outside the Bharathidasan Institute of Management building.
      With the Director and colleagues on campus

  2. 9 – 10 May 2026

    Machine Translation in Bengali: History, Application & AI Frontier

    Resource person, days 3 – 4 of the five-day workshop — Platform Bengali, an arts research project supported by the India Foundation for the Arts, with the Department of English, Ramakrishna Mission Residential College, Narendrapur

    Two days for translators, literature students and digital-humanities researchers on how machines came to find meaning across languages. Day 3: the foundations of NLP; cross-lingual retrieval and the Indian CLIA programme; summarisation and question answering; what BLEU, chrF and COMET measure, and why they struggle with Bengali; and a first look at neural representations. Day 4: the four eras of machine translation — rule-based, statistical, machine-learning hybrids, and neural MT with attention and Transformers, IndicTrans among them — with worked Bengali examples. The other days were taught by Prof. Bidyut Baran Chaudhuri (script, OCR and corpora) and Pritam Dey (LLMs and agents).

    • Dr. Pinaki Bhaskar beside a monk of the Ramakrishna Mission addressing the room, next to the title slide “When machines learned to find meaning across languages”.
      Day 3 opens, Narendrapur
    • Dr. Pinaki Bhaskar speaking, seated beside a screen showing the slide “Machine learning between SMT and NMT”.
      Day 4 — statistical to neural MT
    • The workshop’s participants with Dr. Pinaki Bhaskar, standing together in a lecture hall.
      The participants

Jury & mentoring

  1. August – October 2024

    Samsung Solve for Tomorrow 2024

    Jury member — Solve for Tomorrow, Samsung India's national education and innovation competition for students, run with FITT at IIT Delhi, the Ministry of Electronics & IT and the United Nations in India

    Samsung India’s flagship CSR programme asks school students (14 to 17, on “community and inclusion”) and young people (18 to 22, on “environment and sustainability”) to solve a problem they see around them. In 2024 a hundred shortlisted teams went through regional rounds to twenty, then an innovation walk through Samsung’s offices and R&D centres, and on 30 August pitched at the National Pitch Event at IIT Delhi to a jury of Samsung’s own experts and leaders, which chose the ten finalists. I was one of that jury. What I looked for: innovation that serves a community, solutions that use less energy, whether the team had understood the problem its customers actually have, how far it had carried the idea, and how its members worked together. At the Grand Finale in October the ten teams pitched to a grand jury of Samsung and industry leaders; Eco Tech Innovator from Golaghat, Assam, and METAL from Udupi, Karnataka, were named the champions.

    • Dr. Pinaki Bhaskar in a blue checked blazer on the auditorium stage beside the Samsung Solve for Tomorrow National Pitch Event 2024 standees at IIT Delhi.
      At the National Pitch Event, IIT Delhi
    • Dr. Pinaki Bhaskar, sixth from left, with the other panellists and organisers before the Solve for Tomorrow National Pitch Event 2024 banner at IIT Delhi.
      The jury panel, 30 August 2024
    • Dr. Pinaki Bhaskar in a dark suit on the lit podium in front of the circular Solve for Tomorrow backdrop at the Grand Finale.
      At the Grand Finale, October 2024

Education & service

  1. 2014

    Ph.D. (Engineering)

    Jadavpur University, Kolkata

    Thesis: Answering Questions from Multiple Documents – the Role of Multi-document Summarization. Supervisor: Prof. Sivaji Bandyopadhyay.

  2. 2010

    M.Tech., Computer Technology

    Jadavpur University

    Thesis: A Query Focused Language Independent Multi Document Summarizer and Ad-hoc Monolingual Information Retrieval for English and Bengali.

  3. 2006

    B.Tech., Computer Science & Engineering

    Govt. College of Engineering & Textile Technology, Berhampore (West Bengal University of Technology)

Certificates
  • Neural Networks and Deep Learning, DeepLearning.AI (Andrew Ng) via Coursera, October 2020 Verify on Coursera · Certificate (PDF)
    Coursera course certificate: Neural Networks and Deep Learning, DeepLearning.AI, awarded to Pinaki Bhaskar on 8 October 2020
  • Leadership Communication (CommUP Plus), SRI-B, 2022
  • Introduction to Linguistics, Jadavpur University, 2011
Service
Board of Studies, NIT Meghalaya (2024 – present) · Programme committee: MIKE 2013, INEX 2011 · Reviewer: COLING, CICLing, ICON, NLP-KE, KRAQ, INEX, MIKE (2010 – 2013).
Conference talks
CLEF 2013 (Valencia) · COLING 2012 (Mumbai) · FIRE 2012 (Kolkata) · ENLG 2011 (Nancy) · CLEF 2011 (Amsterdam) · ICON 2010 (Kharagpur) · FIRE 2008 (Kolkata).