Machine Learning Researcher | Ph.D., ISI

Bridging mathematical rigor with
Trustworthy AI systems.

Ph.D. Computer Scientist specializing in Privacy-Preserving Machine Learning, Representation Learning, Trustworthy AI, Information Retrieval, Computer Vision, and Large Language Models.

View Experience Selected Publications
Google Scholar LinkedIn chandan@neuroailabs.in

Core Domains & Expertise

Translating rigorous mathematical representation learning into globally competitive deep-tech systems.

Privacy-Preserving AI

Formulating privacy-preserving representation learning frameworks and deep metric architectures for scalable clustering.

Clinical AI & Health

Leading venture creation of multimodal foundation-model-enabled EMR platforms and robust predictive systems.

Generative EdTech

Scalable LLM-driven personalized learning platforms, automating assessment generation (ExamBoost.ai).

Multimodal LLMs & RAG

Architecting interactive Conversational AI platforms using FAISS and LangChain for querying complex, unstructured multimodal organizational data.

Experience

Lead Data Scientist (Applied Research)

IDEAS - TIH, Indian Statistical Institute 11/2025 - Present
  • Enterprise Conversational AI: Architecting a Retrieval-Augmented Generation (RAG) platform using FAISS and LangChain for interactive querying of complex, unstructured multimodal organizational data.
  • Multimodal Voice-Driven Acquisition: Developing a speech-to-structured-data pipeline integrating ASR, LLMs, and knowledge extraction, reducing manual data entry time by 40%.
  • Asynchronous Video Grounding: Proposing a cross-modal framework to synchronize delayed visual evidence with spoken claims using LLMs and vision encoders.
  • Root Cause Analysis Automation: Formulating ML-driven diagnostics for system failures across multimodal logs, improving anomaly detection rates by 5%.

Co-Founder & Principal AI Architect

NeuroAI Fusion Labs, Kolkata 08/2024 - Present
  • Clinical AI (AICareLens): Leading the venture creation and development of a multimodal foundation-model-enabled EMR platform for personalized clinical decision support.
  • Generative EdTech (ExamBoost.ai): Led the go-to-market strategy and technical deployment for a scalable LLM-driven platform, successfully acquiring 700+ active users.
  • Medical Computer Vision: Architected an occlusion-robust, video-based Freezing of Gait prediction framework using cross-modal deep representation learning and OpenCV.

Data Scientist (Applied Research)

Tatras Data Services, New Delhi 07/2022 - 07/2024
  • Document Layout Analysis: Implemented Table Structure Recognition pipelines utilizing Graph Neural Networks (GNNs) for unstandardized historical documents, increasing parsing accuracy by 5%.
  • Remote Sensing: Architected computer vision pipelines for structural hail damage identification using high-resolution drone imagery, processing 10,000+ images with 92% precision.

Ph.D. Research Fellow

Indian Statistical Institute, Kolkata 07/2016 - 07/2022
  • Formulated a privacy-preserving representation learning framework evaluated on 4 benchmark datasets, outperforming baseline methods by 8-15% clustering accuracy (Q1 pub in Pattern Recognition).
  • Engineered an instance-wise feature attribution framework (L2X) using probabilistic machine learning, improving interpretability while preserving classification performance.
  • Proposed weakly supervised deep metric learning architectures coupled with LSH for scalable, privacy-preserved clustering across 1M+ data points (Q1 pubs in CIKM and IPM).

Education

Ph.D. in Computer Science

07/2016 - 09/2023

Indian Statistical Institute, Kolkata, India

Thesis: Privacy Aware Machine Learning

Supervisors: Dr. Debasis Ganguly & Dr. Ujjwal Bhattacharya

Awarded highly competitive ISI Doctoral Fellowship

M.C.A. (Master of Computer Applications)

07/2008 - 06/2011

Jadavpur University, Kolkata, India

M.Sc. in Mathematics

07/2006 - 06/2008

Jadavpur University, Kolkata, India

B.Sc. (Hons) in Mathematics

07/2003 - 06/2006

University of Kalyani, West Bengal, India

Selected Publications

9 peer-reviewed publications | 130+ Citations

Biswas, C., Ganguly, D., Roy, D., & Bhattacharya, U. (2023). "Weakly supervised deep metric learning on discrete metric spaces for privacy-preserved clustering". Information Processing & Management, 60(1), 103109. [Elsevier, Impact Factor: 8.1, Q1]

Biswas, C., Ganguly, D., Mukherjee, P. S., Bhattacharya, U., & Hou, Y. (2022). "Privacy-aware supervised classification: An informative subspace based multi-objective approach". Pattern Recognition, 122, 108301. [Elsevier, Impact Factor: 9.1, Q1]

Biswas, C., Ganguly, D., Roy, D., & Bhattacharya, U. (2019). "Privacy preserving approximate k-means clustering". Proc. of the 28th ACM International Conference on Information and Knowledge Management (CIKM). [CORE Rank A]

Chaudhury, A., Mukherjee, P. S., Das, S., Biswas, C., & Bhattacharya, U. (2022). "A Deep OCR for Degraded Bangla Documents". ACM Transactions on Asian and Low-Resource Language Information Processing.

Technical Skills

Artificial Intelligence (AI) Large Language Models (LLMs) RAG Computer Vision Document AI & Extraction Privacy-Preserving AI Deep Metric Learning Trustworthy AI & XAI Graph Neural Networks (GNNs) Vision-Language Models (VLMs) PyTorch Transformers / Hugging Face LangChain TensorFlow / Keras OpenCV FAISS / HNSW Python, C, Java AWS / MLOps / Docker