Hi, I am Rishit Kar, a computer engineering undergraduate at DJ Sanghvi College of Engineering, Mumbai University.
Incoming Machine Learning Engineer Intern at IDfy, Asia's leading TrustStack.
Through collaborations with IIT Patna and IIT Mandi, I have worked on explainable medical AI, multimodal learning, and geometric deep learning for engineering systems. I also practice competitive programming and have solved more than 400 problems across Codeforces and LeetCode.
Core Interests:
- Agentic AI & Memory Systems: Persistent state, episodic/semantic retrieval, and long-horizon context architectures
- LLM Reasoning: Multi-step planning, test-time compute, and structured deduction frameworks
- AI Safety & Alignment: Robustness, guardrails, and secure agent tool execution
- Deep Learning: Foundational neural architectures, representation learning, and optimization
Built AI agents including Classifier, Validator, Extractor, and Underwriter for Loan Origination System on Purple Fabric platform, automating manual pipeline across multiple LLMs.
Built multimodal disease detection model on MIMIC (40K+ EHR records, 25 diseases), improving PRAUC by 2.8% via graph architecture.
Optimized propeller geometries via DGCNN on 2048-point clouds, automating preprocessing of 400+ 3D models.
I am most interested in core deep learning, large language models, and agentic AI , their architectures, their theoretical foundations, and their practical applications across vision, language, and reasoning.
Pearl Mody, Mihir Panchal, Rishit Kar, Kiran Bhowmick, Ruhina Karani
ICLR Memagents Workshop, 2026 (Core A*) · Published
Arxiv / Workshop / Code / PyPI /
Neuro-inspired gated memory with utility filtering, episodic buffering, and pruned knowledge-graph consolidation. Evaluated on 100 HotPotQA samples across Qwen, Gemma, and Mistral, achieving F1 0.312 vs 0.068 baseline.
Rishit Kar, Ved Ambulkar, Sargam Nagar, Varun Shenai, Ruhina Karani
PCCDA Conference, 2026 · SCOPUS Indexed Springer LNNS
Code
Multi-modal cancer classifier on the MLOmics benchmark (8,000+ patients, 32 cancer types). Dynamic context gating achieved best precision: GS-BRCA 0.890, GS-COAD 0.947, GS-LGG 0.985.
Cross-platform clipboard manager CLI for developers with persistent history tracking, fuzzy search, and structured content organization. Published as a Python package on PyPI with 5,000+ downloads. Features include history browser, statistics dashboard, and interactive TUI.
Code / PyPI
Tested multiple deep learning models for pneumonia classification on 5,000+ chest X-ray images. Implemented a multimodal pipeline integrating spatial attention features with a Graph Attention Network (GAT). Custom CNN achieved 7–9% higher accuracy over ResNet and GNN baselines.
Code
Reliable transport protocol with sliding-window flow control, retransmission, packet ordering, and SHA-256 integrity checks.
Code
Workshop · IDfy · Mumpy
Event / Slides
Presented in front of 75 people including students and professionals. An introductory workshop on PySpark, distributed data processing, and practical big-data workflows.
Webinar · ACM College Committee
Video / Slides
A talk on finding research opportunities, writing effective outreach, and building a student research profile.
Dwarkadas Jivanlal Sanghvi College of Engineering, Mumbai · 2023 – 2027 · CGPA 9.14
Research Head at DJS ACM
Mentored 15+ students in Machine Learning, facilitating research projects and collaborating with college professors to advance ML research within the community.