All Projects
A comprehensive collection of research projects, hackathon wins, and technical innovations spanning neuro-AI, computer vision, and computational systems.
Featured Projects
HiCL - Hippocampal Inspired Continual Learning
AAAI 2026Developed HiCL, a hippocampal-inspired architecture for continual learning that significantly reduces forgetting in neural networks. Engineered a DG-Gated MoE system that dynamically routes tasks to specialized subnetworks without a separate gating network. Achieved near SOTA performance on continual learning benchmarks, while reducing computational costs compared to other methods.
Blindspot
Bitcamp 2023 - 1st Place WinnerCreated a prototype of image-recognition glasses for the visually impaired. Applied YOLO object detection algorithm by fine-tuning its parameters, optimizing the data transfer and processing pipeline to minimize latency. Integrated the hardware with software and created the facial recognition algorithm.
ThoughtWheels
TreeHacks 2024 - Top 10/500 ProjectsCreated a smart wheelchair for people with severe mobility impairment to move a wheelchair using their Brain Waves. Designed signal processing algorithms to interpret EEG signals from Muse 2, developed a robotic control system for precise wheelchair navigation based on micro facial movements and built an algorithm to use for a bigger sensor and imagined movements.
MediChain
HackMIT 2024 - Best Blockchain HackDeveloped a blockchain-based platform for secure and transparent management of patient data, enhancing privacy and data integrity. Integrated blockchain technology with Neo4j knowledge graphs to create a decentralized ledger system for tamper-proof storage and authorized access to medical records, hosted and launched a decentralized storage system with these records to stop server shortage.
Other Projects
TrashPro
Sustainable AI InnovationSustainable waste management app powered by cutting-edge AI that sorts trash into Recycle, Compost, or Landfill categories, promoting eco-friendly habits for a greener planet. Uses computer vision and machine learning to identify waste types in real-time.
DeceptionTime
ICDC 2021 - Published in SpringerDesigned a large-scale deep learning model to predict stock price movement and market volatility using a sophisticated model based on transpose CNNs and CNNs, ESNs, and GRUs. Architecture surpassed most State-of-the-Art models, including Google's inception, beating models on benchmarks by 10 to 200%.
Generative Underwater Imaging
ICRA 2026Worked on a diffusion-based model to generate underwater images from backscattering, illumination and text descriptions of scenes. Finetuned Atlantis to work with this project, creating realistic underwater scenes from SONAR emissions and text descriptions.
Avellaneda-Stoikov with Dentate Gyrus
Research ProjectDeveloping an improved version of the Avellaneda-Stoikov Market Making model by adding a layer of pattern separation inspired by the Dentate Gyrus neural circuit. Introduces predictive modeling to traditional market making strategies.
AlphaSynuclein to Parkinson's Predictor
PreprintDeveloped a machine learning model to predict Parkinson's disease progression from AlphaSynuclein protein patterns. Combines molecular biology insights with deep learning for early disease detection.
Jira Ticket Assistant
Marriott InternationalStreamlined productivity of the Loyalty IT team through a Jira Ticket assistant, running searches on past Jira tickets and providing recommendations to solve tickets. Increased team productivity by 20% and reduced ticket resolution time by 50%.