| ||||||
Aishik Konwer, PhDAdvanced AI Research Scientist Center of Advanced AI, Accenture Email: akonwer@cs.stonybrook.edu, Mobile: +1-6317471244 Resume • Google Scholar • Github • Linkedin | ||||||
At the Center of Advanced AI, Accenture, I build multimodal generative agents for video synthesis, develop LLM post-training pipelines with RL-based optimization (GRPO), and engineer spatially grounded VLM pipelines for anomaly detection and zero-shot object recognition. I contribute to multimodal agentic capabilities for AI Refinery and am currently exploring multi-view VLA representations for embodied AI and robotics.
I completed my PhD in Computer Science from Stony Brook University (April 2025). My thesis focused on designing algorithms that learn efficiently from imperfect 2D/3D/multi-stained medical imaging datasets, spanning segmentation, caption generation, and future timepoint prediction. My work includes enhancing multimodal foundational models (VLMs) with efficient prompts and preference optimization, applying GANs and conditional diffusion models, and customizing meta-learning and few-shot algorithms for clinical tasks.
Research Interests: Multimodal LLMs (VLM/VLA), LLM Agents, Generative AI (Diffusion Models, Video Generation), RLHF & Post-Training (SFT, DPO, GRPO), Embodied AI, Video Understanding, Meta-learning, Few-shot Learning, Domain Generalization.
I am currently on the job market for Research Scientist roles.
Recent News
Education
| ||||||
| Stony Brook University, USA
PhD in Computer Science & Engineering Aug 2019 - Apr 2025 | ||||||
| ||||||
| Institute of Engineering & Management, India
Btech in Electronics and Communication Engineering Aug 2013 - June 2017 | ||||||
Research Experiences
| ||||||
| Indian Statistical Institute, Kolkata, India
Under Prof. Umapada Pal Sept 2016 - Sept 2017 [Certificate] | ||||||
| ||||||
| Indian Institute of Technology Roorkee, India
Under Prof. Partha Pratim Roy Sept 2017 - March 2018 [Certificate] | ||||||
Industry
| ||||||
| Accenture, USA
Advanced AI Research Scientist May 2025 - Present | ||||||
| ||||||
| GE Healthcare, USA
AI Scientist Intern May 2024 - August 2024 | ||||||
| ||||||
| SRI International, USA
Deep Learning Research Intern May 2023 - August 2023 | ||||||
| ||||||
| Roche, USA
Advanced ML Research Intern May 2022 - August 2022 | ||||||
| ||||||
| Cognizant, India
Datawarehouse developer Dec 2017 - July 2019 | ||||||
Selected Research
|
Reasoning-Guided Grounding: Elevating Video Anomaly Detection through Multimodal Large Language Models
|
|
Gaze2Report: Radiology Report Generation via Visual-Gaze Prompt Tuning of LLMs
|
|
Physical AI: The Next Frontier in AI and Robotics to Build Truly Autonomous Machines
|
|
Enhancing SAM with Efficient Prompting and Preference Optimization for Semi-supervised Medical Image Segmentation
|
|
MetaStain: Stain-generalizable Meta-learning for Cell Segmentation and Classification with Limited Exemplars
|
|
Enhancing Modality-Agnostic Representations via Meta-Learning for Brain Tumor Segmentation
|
|
Temporal Context Matters: Enhancing Single Image Prediction with Disease Progression Representations
|
|
Attention-based Multi-scale Gated Recurrent Encoder with Novel Correlation Loss for COVID-19 Progression Prediction
|
|
Predicting COVID-19 Lung Infiltrate Progression on Chest Radiographs Using Spatio-temporal LSTM based Encoder-Decoder Network
|
|
|
|
|
|
|
|
|
|
|
|
A New GVF Arrow Pattern for Character Segmentation from Double Line License Plate Images
|