.png&w=256&q=75)
🤓 Latest Submissions
.png&w=828&q=75)
VOLMIND — Autonomous Options Research Terminal
VOLMIND trades the divergence between what an AI believes and what the options market is pricing in. A LangGraph pipeline of specialized agents — News, Fundamental, and Probability — each call an LLM to reason over live Alpaca market and news data and produce an evidence-required probability estimate; unsupported forecasts have their confidence automatically clamped. That estimate is compared against the market-implied probability computed directly from the live Alpaca option chain. A second, independent LLM pass — the Diligence Agent — stress-tests the resulting thesis for thin evidence, IV/liquidity red flags, and whether the edge clears configured thresholds, halting the pipeline before a trade is ever proposed if it doesn't hold up. Only cleared theses reach the Risk Agent, which enforces per-trade loss limits, a portfolio-wide position cap, and a daily loss circuit breaker before a real paper order is submitted. Execution isn't the finish line: a Position Monitor marks every open position to market and closes it on a take-profit, stop-loss, or days-to-expiry rule, and agent reputation is scored against real trade outcomes. Left running, VOLMIND scans its watchlist and manages its book autonomously — market-hours-gated, fail-closed to paper trading — for the length of a competition, with every decision logged with its full reasoning for an auditable trail. Built as a hackathon agent, but the same pipeline doubles as a decision-support terminal a desk could run for research and alerts, with auto-execution simply switched off.
4 Sep 2026
.png&w=828&q=75)
SOAP Copilot: AI Clinical Scribe on AMD
Physicians spend roughly two hours a day on clinical documentation, time pulled directly away from patient care, and a well-documented driver of burnout that costs the US healthcare system an estimated $8.3B a year in lost productivity. SOAP Copilot is a multi-agent AI system that turns a raw doctor-patient conversation into complete, structured clinical documentation in under 30 seconds. Three specialized agents, all built on Llama 3.3 70B Instruct, handle the pipeline in sequence: a SOAP Generator produces a structured Subjective/Objective/Assessment/Plan note, an ICD Coder extracts diagnosis codes with confidence scores, and a Summary Writer rewrites the note as a plain-language patient summary. The entire stack runs on open-source models, so no PHI ever needs to leave the user's own infrastructure. The project was originally prototyped and LoRA fine-tuned on a single AMD MI300X (192GB HBM3) via AMD Developer Cloud, running vLLM on ROCm 7.2. That memory headroom is what makes serving a 70B-parameter model in BF16 precision practical on one chip, something no consumer GPU and few cloud GPUs can do. For this submission, production inference is served through Fireworks AI, which itself runs on AMD Instinct hardware, so the project demonstrates the AMD AI stack end to end: self-managed fine-tuning through to hosted inference. Roadmap items include a distilled specialist model for lower-cost deployment, retrieval-augmented generation over clinical guidelines using the Qdrant vector database already wired into the architecture, and an on-premise, HIPAA-compliant deployment path for health systems that cannot send patient data to any external API.
13 Jul 2026
