i'm an ai/ml engineer working at the intersection of model training, red teaming, and open-source tooling.
day to day: fine-tuning and quantizing language models, running red teaming and refusal-direction work, and building tools around them. local-first models when they make sense, hosted apis when they don't.
research that ships. evaluations that don't lie. tools other engineers actually use.
ML framework for fine-tuning SLMs via Domain-Routed On-Policy Iterative Correction. Combines verified repair, delta-span subtraction, and ZPD-weighted curriculum scheduling. L = L_self + λ_r L_repair + λ_delta L_delta, where w_zpd = 4·p̃·(1−p̃) and p̃ = (s+0.5)/(K+1).
Multi-agent debate and vote coordination system on Solana blockchain. Agents assume distinct personas to debate complex questions, with full session transcripts hashed and recorded on-chain for verifiable AI consensus.
Autonomous triple-AI engine that analyzes codebases and opens validated PRs hourly with cross-repo global memory.
quantized gguf builds of ZAYA1-8B-Coder for local inference via llama.cpp, ollama, and lm studio.
HFquantized gguf builds of gemma-4-E4B-it-Coder for local inference via llama.cpp, ollama and lm studio.
HFfine-tuned Hcompany Holo-3.1-9B for coding tasks. merged model optimized for python and software development.
HFquantized gguf builds of Holo-3.1-4B-Coder for local inference via llama.cpp, ollama and lm studio.
HFquantized gguf version of qwopus 9b unfettered for efficient local inference. 100% compliance on harmful evals.
HFquantized gguf builds of Fara-7B-Abliterated-v2 for local inference via llama.cpp, ollama, and lm studio.
HFsurgical unalignment of Qwen 0.5B Instruct optimized for low-end hardware. 100% compliance, zero refusal. runs on 1GB RAM.
HF31,909 rows of curated code/math post-training data derived from OpenBMB UltraData. includes SFT and think splits.
HF9B uncensored language model. directional ablation applied to remove refusal mechanisms. 100% compliance on harmful evals.
HFfine-tuned Hcompany Holo-3.1-4B for coding tasks. merged model optimized for python and software development.
HFfine-tuned google gemma-4-E4B-it for coding tasks. multimodal model optimized for code generation.
HFmixed-origin public coding data with 2,700+ prompt/response pairs for llm training experiments.
HFrefusal-direction-orthogonalized variant of microsoft/Fara-7B. 98.75% compliance on held-out harmful evals.
HFmerged coder model from Zyphra/ZAYA1-8B plus custom lora. +24% lift on python code evaluation gate.
HF200+ curated refusal-compliance prompt pairs for red teaming and adversarial evaluation.
HFlora adapter for google gemma-4-E4B-it focused on code generation and software development tasks.
HFlora/qlora adapter for Holo-3.1-4B focused on coding and python development.
HFlora adapter for Zyphra/ZAYA1-8B focused on python code generation. +101% relative lift over base.
HFquantized gguf builds of Mellum2-12B ablated for refusal removal. MoE architecture with per-expert per-layer projected ablation.
HFabliterated Mellum2-12B thinking model from JetBrains. refusal-direction orthogonalized with CoT steering for reasoning tasks.
HFquantized gguf builds of LFM2.5-8B-A1B Coder for local inference via llama.cpp, ollama and lm studio.
HFfine-tuned LiquidAI LFM2.5-8B-A1B MoE model for real-world coding tasks. multilingual and conversation-optimized.
HFlightweight lora adapter for LFM2.5-8B-A1B focused on real-world coding and multilingual tasks.
HFfine-tuned sapientinc HRM-Text-1B for code generation. trained on HumanEval and MBPP benchmarks.
HFlora adapter for HRM-Text-1B focused on python code generation and coding benchmarks.
HFlora adapter for Qwen2.5-7B optimized for agentic tasks and tool-use workflows.
HFexperimental lora adapter for Qwen2.5-Coder-3B exploring cheat/control dynamics in small language models.
HFexperimental lora adapter for Qwen2.5-Coder-3B exploring clean baseline dynamics in small language models.
HF