AI Researcher · ML Engineer
Open to research & applied-ML rolesDivyansh Pandey
Turning research into real systems.
ML engineer and researcher working on Neurosymbolic AI, knowledge graphs, sparse autoencoders & mechanistic interpretability, VLMs & LVLMs, embodied AI, federated learning, and LLM fine-tuning.
Currently at IAIRO, IISER Kolkata, and Pragya Lab (BITS Goa). 3× Springer Nature published author — open to research collaborations and applied-ML roles.
Experience
AI Research Intern
· InternshipIndian AI Research Organization (IAIRO)
Mar 2026 – Present · Remote
- Researching Neurosymbolic AI and knowledge integration — frameworks that fuse neural networks with symbolic reasoning for interpretable, generalizable systems.
- Supervisors: Dr. Amit Sheth, Dr. Vedant Khandelwal
AI Research Associate
· ApprenticeshipIISER Kolkata
Mar 2026 – Present · Remote
- Researching Sparse Autoencoders (SAEs) and mechanistic interpretability — decomposing LLM internals to understand feature representations and decision-making.
NLP Research Intern
· InternshipPragya Lab, BITS Goa
Jun 2026 – Present
- Research in NLP and LLM evaluation — building systematic methodologies to assess large language model behavior and outputs.
- Supervisor: Dr. Amitava Das
Machine Learning Researcher
· Full-timeManipal University Jaipur
Jan 2026 – Present · Jaipur, Rajasthan · On-site
- Applied ML research in computer vision and deep learning with PyTorch and medical imaging datasets.
- Developing and evaluating model architectures for classification and segmentation on real-world data.
ML Research & Development Intern
· InternshipVIGIL Labs, IIT Hyderabad
Apr 2025 – Jul 2025 · Hyderabad, Telangana · Remote
- Engineered a decentralized Federated Learning model for medical image classification & segmentation, surpassing baseline test accuracy by 20% on complex non-IID real-world data.
- Reduced global communication round time by 45% and used 55% fewer resources than baselines by optimizing communication protocols.
- Orchestrated secure ML workflows addressing data heterogeneity and strict distributed-data-privacy requirements.
Publications
View researchDEXNet: An Ensemble Model Integrating DenseNet, EfficientNetB3, and XGBoost for Histopathological Lung and Colon Cancer Classification
Ensemble architecture combining DenseNet and EfficientNetB3 for hierarchical feature extraction with XGBoost classification, augmented by Class-Selective Image Preprocessing (CSIP), Grad-CAM interpretability, and a HIPAA/GDPR-compliant AWS deployment pipeline with federated learning provisions for multi-institutional privacy-preserving inference.
FedBound: A Boundary-Aware Optimization Strategy for Federated Medical Image Segmentation Under Non-IID Data
A lightweight boundary-aware optimization strategy for federated medical image segmentation that emphasizes contour pixels during local training without adding communication overhead. Evaluated across six segmentation architectures on the ISIC 2018 skin-lesion dataset under a Dirichlet non-IID partition of 100 federated clients — improving boundary precision (HD95) and reducing cross-client variance while preserving Dice/IoU performance.
Barbell Exercise Classification and Repetition Counting
Engineered a robust ML system for barbell exercise classification and repetition counting using MetaMotion sensor data, achieving over 90% accuracy through comprehensive feature engineering and outlier detection pipelines for precise human activity recognition.
Projects
View allRAGineer
RAG · NLPProduction-ready local RAG Text-to-SQL chatbot that converts natural language into safe PostgreSQL queries with 84% SQL accuracy.
Real-Time Fraud Detection
MLOps · ProductionHigh-performance FastAPI microservice achieving 90% latency reduction (50ms → 5.4ms) with full production observability.
GetAnime
RAG · GenAIRAG-powered semantic anime recommendation engine with sub-second latency and 95% relevance accuracy across 12,000+ entries.
Currently exploring
- Neurosymbolic AI & Knowledge GraphsFusing neural networks with symbolic reasoning and structured knowledge for systems that generalize and stay interpretable.
- Sparse Autoencoders & Mechanistic InterpretabilityDecomposing LLM internals to understand how models represent features and make decisions.
- VLMs, LVLMs & Embodied AIVision-language models and grounded agents for real-world perception and action.
- Efficient Inference & LLM DeploymentModel optimization and serving language models at scale.
Skills & Stack
Languages
ML / DL
LLMs & Agents
MLOps & Cloud
Databases
Frameworks & Tools
Honors & Awards
3× Springer Nature Published Author
· 2024 – 2026SN Computer Science (Q1) · Discover Computing (Q2) · ICDEC 2024
Peer-reviewed research in ensemble deep learning, federated medical image segmentation, and sensor-based human activity recognition.
Dean's List — Excellence in Academics (Highest GPA)
· 2024Manipal University Jaipur
Awarded for the highest Grade Point Average in Computer Science and Engineering (AI & ML).
2× Dean's List — Excellence in Off-campus Achievements
· 2025Manipal University Jaipur
Recognition for research publications and competition results beyond the classroom.
Runner Up — Xiaomi Ode2Code 3.0
· 2023Xiaomi India
National-level coding competition organized by Xiaomi India.
Positions of Responsibility & Volunteering
Member · Open Source Contributor
· Part-timeThe Red Wood Lab
Jan 2025 – May 2025 · Jaipur, Rajasthan · Remote
- Contributed a multivariate Ridge regression model for electrical energy prediction on Micro Gas Turbine data (achieving R² = 0.9785 on test set).
- Worked across ML, NLP, and data science projects within the lab's open research ecosystem.
Open Source Contributor
· Part-timeSocial Winter of Code (SWOC)
Jan 2025 – Mar 2025 · Remote
- Contributed to open-source projects spanning LLMs, Computer Vision, and NLP as part of the SWOC program.
Open Source Contributor
· Part-timeHacktoberfest
Oct 2024 · Remote
- Contributed ML and data science notebooks to open-source repositories during Hacktoberfest 2024. Earned Level 1 badge.
Software Engineering Virtual Simulation
· TraineeJPMorgan Chase & Co. (via Forage)
Jan 2024 · Remote
- Completed a simulation focused on engineering tasks in JPMorgan Chase's credit-card rewards department.
- Created a new class to get an existing system up and running; authored a comprehensive unit test suite in Java.
Education
Manipal University Jaipur
B.Tech (Hons.) Computer Science Engineering — Major in AI & ML
Graduated · Sept 2022 – Jun 2026 · CGPA: 8.53
AI, Machine Learning, Deep Learning, Computer Vision, NLP
- 3× Dean's List — Excellence in Academics (Highest GPA) & Off-campus Achievements
- 3× Springer Nature Published Author — SN Computer Science (Q1), Discover Computing (Q2), ICDEC 2024
- Runner-Up — Xiaomi Ode2Code 3.0 national coding competition