Cade Miller@cademiller23280
2
Events attended
2
Submissions made
United States
9+ years of experience
About me
Hi, My name is Cade I have curated multi agent systems that have performed at frontier level with reaching 70.2% on OSWorld-verified and SWE bench verified performed 80% while SWE Bench Pro 56%. As well as curated with my team Mixture of Experts with personalized models for different domains with a full orchestration exceeding 700B parameters.
🤝 Top Collaborators
.png&w=640&q=75)
.png&w=640&q=75)
🤓 Latest Submissions
THESIS: options volatility engine
THE THESIS Options routinely trade away from fair value. Those gaps look like free money and usually aren't — the spread required to capture one is typically larger than the gap. Most strategies never verify this and bleed out slowly. THESIS treats verification as the primary task. AI LOGIC 1. CAPTURE — the full chain is paged off the wire and written to disk byte-for-byte, so any conclusion is re-derivable from exactly what the market said. 2. SURFACE — each expiry gets an SVI smile. Butterfly arbitrage isn't checked post-hoc; it's made structurally impossible inside the objective. Slices fit sequentially so calendar monotonicity is structural. Every fit carries a domain of validity; outside it the model declines to have an opinion. 3. DETECTION — a contract is dislocated when its price differs from the surface by more than the spread and fit error can jointly explain. Significance is measured in noise, not size. 4. STRUCTURE — each signal maps to a defined-risk vertical. The hedge leg must be tightly quoted AND on-model; otherwise a named refusal. RISK GATES Five vetoes: fit confidence · clean partner · measured cost (edge must clear 1.5x a $2.00 round trip, MEASURED by opening and closing a real spread) · wide-quote artifact (at high vega a $4 spread looks tight in vol and still costs $4) · risk-adjusted return. Beneath them, hard limits: per-trade cap, total open risk, concurrency, daily stop. Defined-risk only; max loss known before transmission. ALPACA Multi-leg option orders, trade-update WebSocket (the paper stream sends JSON in binary frames), snapshots/contracts with exhaustive pagination, and an MCP server exposing analyze surface, find_dislocations, check viability. Lifecycle proven live: open → fill → position → close → flat, zero residue. RESULTS Every layer states what it won't do and stops. The refusals are the product. github.com/Cademiller23/Thesis · account PA3XUA5XFEOO
4 Sep 2026
TELMED AI DOCTOR
Telmed AI Doctor is an AI-powered telemedicine platform built to close Africa's healthcare access crisis. Nigeria has only 0.4 doctors per 1,000 people. Millions suffer from treatable conditions simply because qualified medical guidance is out of reach. Telmed deploys AI as the first line of medical response — available 24/7, on any device, at zero cost to the patient. Users interact via text, voice, or image. They describe symptoms naturally or photograph affected areas and receive a structured clinical interpretation, severity classification — Low, Moderate, or Urgent — a personalised medication schedule, and clear escalation triggers for when a human doctor is immediately required. The AI engine runs on a multi-model cross-validation architecture. Google Gemini handles primary diagnostic reasoning while DeepSeek independently validates the same input — eliminating hallucination risk on clinical outputs. Google Cloud Vision powers image-based symptom analysis. Google Places connects users to the nearest verified pharmacy. Built on Node.js and Express.js with a vanilla JavaScript frontend optimised for low-bandwidth African mobile networks. Recognised as a Top 100 project out of 3,000+ submissions at the lablab.ai AI Agent Olympics Hackathon, May 2026. Telmed is not a product. It is infrastructure. The kind Africa should have built a decade ago.
19 Jun 2026
