Projects

A collection of systems I've built, ranging from on-device AI and LLM agents to computer vision and generative models.

SpikeData

Python-based module for neural data analysis pipelines

Role: Developer

TL;DR

  • Flexible and robust module for creating spike sorting pipelines.
  • Simplifies analysis of intrinsic neural recordings across large-scale HD-MEA datasets.
  • Part of the Integrated Analysis Toolkit used by the Braingeneers group.

System Architecture

Raw Electrophysiology → Signal Processing → Spike Sorting → Analysis

Results & Validation

Enabled scalable analysis for 1000+ channel neural recordings.

Tech Stack

PythonNeural Data ScienceSoftware Engineering

Polaris

On-device AI for preventative maintenance for fishermen

Role: Lead ML Engineer

TL;DR

  • Built an on-device AI system to monitor fishing equipment and predict maintenance needs.
  • Won 1st place at Sushi Hackathon 2025.
  • Optimized for low-power edge deployment on custom hardware.

System Architecture

Sensors → Edge AI → Real-time Maintenance Alerts

Results & Validation

1st Place Winner at Sushi Hackathon (2025).

Tech Stack

Edge AIIoTPythonSensor Fusion

Sushi Agent

Autonomous AI agents for coupon discovery

Role: Lead ML Engineer

TL;DR

  • Developed agents that crawl and negotiate for private discount codes and coupons.
  • Utilized LLM-based decision making for navigating complex web interfaces.
  • Implemented automated validation and verification of discovered codes.

System Architecture

LLM Agents → Web Crawling → Coupon Validation

Results & Validation

2nd Place Winner at Sushi Hackathon (2024).

Tech Stack

LLMAgentsPythonNLP

StatScanner

Computer Vision for table digitization

Role: Lead Developer

TL;DR

  • iOS app that converts printed tables into interactive visualizations.
  • Custom OCR pipeline optimized for tabular data structure.
  • Real-time data visualization and CSV export functionality.

System Architecture

Camera → OCR → Table Parsing → CoreML Visualization

Results & Validation

Shipped to App Store; processed 10k+ scans with high structural accuracy.

Tech Stack

SwiftCoreMLComputer VisioniOS

Rubauto

Diffusion model for stylistic music generation

Role: Project Lead

TL;DR

  • Built a generative model for creating music in specific artistic styles.
  • Experimented with diffusion techniques for high-fidelity audio synthesis.
  • Developed a custom training pipeline for style-specific music data.

System Architecture

Audio Data → Diffusion Model → Stylized Audio Output

Results & Validation

Successfully generated high-fidelity stylistic variations of input tracks.

Tech Stack

Diffusion ModelsPyTorchAudio MLGenerative AI

Gesture-Driven AR Navigation

CV-assisted spatial interaction

Role: Co-Developer

TL;DR

  • Created an augmented reality navigation system controlled by hand gestures.
  • Implemented custom gesture recognition and mapping to 3D space.
  • Built a low-latency AR overlay for intuitive spatial navigation.

System Architecture

Camera → CV Pipeline → Gesture Mapping → AR Overlay

Results & Validation

1st Place at Helix Hacks II (2021).

Tech Stack

Computer VisionAR/VRPythonUnity

Project Tourist

NLP-based personalized travel planner

Role: Co-Developer

TL;DR

  • Natural language interface for generating optimized travel itineraries.
  • Integrated multiple APIs for real-time travel and lodging recommendations.
  • Built a constraint-aware engine for personalized trip planning.

System Architecture

NLP Engine → Recommendation Engine → Trip Generation

Results & Validation

Provided users with optimized, personalized, and constraint-aware travel plans.

Tech Stack

NLPPythonAPI Integration

Link Saver

Academic productivity launchpad

Role: Co-Developer

TL;DR

  • Productivity tool for students to manage and quickly launch academic links.
  • Designed for high-speed access to meetings, assignments, and resources.
  • Simple, intuitive interface for organizing complex student workflows.

System Architecture

Web App → Link Management → Quick Access UI

Results & Validation

3rd Place Winner at BAYCO Hackathon (2021).

Tech Stack

ProductivityWebJavaScriptUX

Kam Show Podcast

Science & engineering communication

Role: Host & Producer

TL;DR

  • Produced and hosted a podcast exploring cutting-edge science and tech.
  • Featured interviews with researchers and engineers across various fields.
  • Managed end-to-end production, from booking to audio engineering.

System Architecture

Interviews → Audio Engineering → Distribution Platforms

Results & Validation

Reaching listeners globally with deep-dives into emerging technology.

Tech Stack

Content CreationAudio EngineeringCommunication