Computer Science student exploring how intelligent systems reason, plan, and verify — not just generate more tokens.
I'm a Computer Science student interested in AI reasoning, research engineering, and reliable agentic systems. I build research tools, interactive experiments, and software workflows that make complex ideas easier to inspect, test, and reproduce.
My focus is on evidence over appearance, verification over assumption, and useful systems over impressive demos.
A map of my public work across AI research, agent engineering, specialist toolkits, and applied systems. Each project name opens its repository.
- Argumentum Research Orchestra — A self-hosted, multi-agent research infrastructure for literature discovery, evidence synthesis, methodological critique, and reproducible manuscript workflows.
- ReasoningAI Studio — An evidence-first research toolkit with portable skills, reusable prompts, templates, and local utilities for reproducible research.
- Neural Networks Playground — An interactive machine-learning lab for training small neural networks and inspecting activations, gradients, optimizers, and decision boundaries.
- Algorithm DSA Forge (Archived) — A visual learning environment for data structures, algorithms, pathfinding, machine learning, and AI reasoning, with step-by-step execution.
- AI Agent Engineering Toolkit — A portable library of agent skills, specialist roles, and workflow templates for software engineering, research, testing, and technical delivery.
- LaunchCrew — A Codex-first, spec-driven toolkit for turning product ideas into scoped plans, delegated implementation, verified changes, and clear handoffs.
- Claude Garage — A Claude Code plugin that brings structured startup-style workflows to planning, parallel building, testing, auditing, and shipping.
- Implicit State Machines — Agent workflows for finding hidden application lifecycles and refactoring fragile, boolean-heavy state into explicit, testable models.
- AI Creative Production Toolkit — A library of AI agent skills and local Python tools for image, video, audio, 3D, game art, and creative production workflows.
- JaneStreet Agentic Skills — Specialist playbooks for quantitative trading, market microstructure, order-book systems, low-latency engineering, risk controls, and research validation.
- KINETIQ Machine Simulation Skills — Portable skills for interactive 3D machine explainers, mechanical simulation, physics validation, and offline delivery.
- Pixelcraft — A specialized skill library for identifying and improving generic interfaces, visual hierarchy, typography, layout, accessibility, and design consistency.
- LinkedIn Skills 360 — A model-agnostic AI skill ecosystem for LinkedIn workflows across careers, networking, content creation, sales, recruiting, and hiring.
- MarketPilot AI — A model-agnostic trading research toolkit for market scans, risk checks, scenario replay, walk-forward evaluation, and paper-only workflows.
- The Sharks — A multi-agent startup evaluation framework that pressure-tests customers, revenue, unit economics, competition, and founder-market fit.
- Kinoquence — A local-first, model-agnostic production workflow for animated shorts, with structured storyboards, asset tracking, captions, credits, and quality checks.
- SheerFlow — SheerID Workflow Studio — A JavaScript workbench for authorized SheerID API integrations, verification workflows, document uploads, and auditable job tracking.
- HoneyPot Guard — A local-first pre-commit scanner for leaked secrets, suspicious outbound endpoints, and risky code patterns.
- MAPIE — A repository copy of the upstream MAPIE project, a scikit-learn-compatible library for conformal prediction and risk control. Listed as an ecosystem reference, not as original authorship.
Languages and Runtime
AI, Machine Learning, and Research
Agentic Development
Engineering and Workflow


