Engineering artifacts
| No. | Code | Project | Record |
|---|---|---|---|
| 01 | SAFE | Model safety | Contract |
| 02 | NLP | NLP research | Published |
| 03 | MATL | Bioplastic formulation | Pending |
| 04 | QUANT | Algorithmic trading | Supplementary |
| 05 | DRON | Dronova | Deep tech |
| 06 | LRNF | LearnefAI | Edtech |
Contract AI safety researcher and red-teamer. Passed technical screening and secured paid work testing frontier models; identified universal jailbreak-style vulnerabilities and helped drive mitigations pre-deployment.
LLM safety · adversarial testing · evaluation
Peer-reviewed NLP work on multi-persona interaction as a prompting method to improve logical reasoning in LLMs (vs. standard chain-of-thought baselines).
“Town Hall Debate Prompting: Enhancing Logical Reasoning in LLMs through Multi-Persona Interaction.”
NLP · prompting · evaluation
Independently designed and engineered a novel bioplastic formulation and filed a full utility patent with the USPTO; currently responding to office action.
Materials engineering · experimentation · patent prosecution
Built and backtested systematic equity strategies in Python (momentum-driven signals, asymmetric risk construction), plus an end-to-end research workflow for iteration and evaluation.
Python · backtesting · statistics
Building the software foundations for autonomous, self-improving drone swarms. The current focus is simulation and evaluation harnesses for coordinating how drones plan, map, and move together.
That work spans terrain-aware surveying, obstacle navigation, and control under separation and actuator constraints, tested in Python and PX4 software-in-the-loop. The longer-term direction is to carry that autonomy into purpose-built hardware architectures.
Drone autonomy · Python · PX4 simulation
An AI-powered learning platform that builds a structured course around what you want to learn. It brings lessons, quizzes, and flashcards into one self-paced path, with progress that carries from topic to topic.
Instead of choosing from a fixed course catalog, learners start with a subject or question and generate a sequence of chapters, explanations, and practice. Built for middle and high school learners ages 13+, with XP, streaks, and achievements to support a consistent study habit.
AI-generated courses · self-paced learning · progress tracking
Research workflow. No performance or investment-return data is displayed.