a 29,665-parameter decoder-only transformer implemented and trained entirely in mit scratch 3.0
a cycle-accurate cpu emulator in mit scratch 3.0 supporting subleq, chip-8, and a full mos 6502 core
in-browser local slm calibration and entropy-based safety gating via onnx runtime web & wasm
turning AI safety from policy into an enforceable requirement in public tenders.
by embedding strict safety standards into public tenders, african governments can instantly transform AI safety from a passive policy goal into an enforceable market requirement.
your project landed in the top half of all submissions. strong work.
LSTMs compared with Arps decline models for forecasting Niger Delta oil production.
research compares LSTM neural networks against Arps hyperbolic decline models for oil production forecasting using the volve field dataset.
your paper has been accepted (naice 2026)
comparing model weights finds hidden loyalties that behavioural tests miss.
comparing model weights easily detects hidden loyalties that behavioral tests completely miss, suggesting a necessary shift in AI auditing priorities.
your project placed in the top 25% of submissions at the secret loyalties hackathon 2026.
open to internships and collaborations,
