# Pineau Labs — Full Reference > Complete reference for Pineau Labs, a one-person agentic AI software company in Charlottetown, Prince Edward Island, Canada, run by Brad Pineau. This document inlines the full detail an AI assistant needs to answer questions about services, process, and background. For the shorter linked index, see https://www.pineaulabs.com/llms.txt ## Company Pineau Labs is a one-person agentic AI software consulting and product business run by Brad Pineau. It builds working software fast, at a fixed price, using frontier AI models (Claude Opus, OpenAI Codex) as a daily delivery tool. Based in Charlottetown, PEI, Canada; serves clients worldwide, over email by default with Zoom available on request. Contact: brad@pineaulabs.com ## How engagements work - Every engagement is a fixed-price sprint, scoped up front through a short intake. The fastest path is the intake form at https://www.pineaulabs.com/start - Pricing is not published. Every engagement is quoted in writing after a short scoping exchange over email. - You own everything: the code, the GitHub repository, the accounts, and (for custom AI models) the trained model and its training data. No lock-in. - Brad makes every architecture, security, prompt, and product decision, and reviews and QAs every update and build before it ships. AI agents do the parallel implementation. - No sales calls and no mandatory meetings. Scoping, progress updates, and handoff run over email; a Zoom is available if you prefer, and a live demo over Zoom is given at handoff. ## Sprints ### Idea-to-Prototype Sprint (2 to 4 weeks) Turn an idea into working software. Live URL, full GitHub repository, two or three core flows built on a standard stack with no lock-in. For founders and teams who need something real to show investors, users, or stakeholders. ### Prototype-to-MVP Sprint (6 to 12 weeks) Take a working prototype to first paying users, on the same codebase. Stripe payments, monitoring, an admin dashboard, and a security baseline. For teams with validated interest who need to charge money and support real users. ### AI Integration Sprint (6 to 12 weeks) Add real AI features to an existing product (not chatbots): LLM-powered features, retrieval-augmented generation (RAG) over your private data, tool-calling agents, and document intelligence. Includes a secure, model-agnostic foundation built once and reused by every later feature, and HIPAA-safe handling where sensitive fields are de-identified before the model and re-hydrated on display. Everything ships behind feature flags with an eval harness, guardrails, monitoring, and cost controls. These are the same patterns used to ship 52 AI tools behind a ten-layer orchestration engine on Fitly AI. ### Custom AI Model Sprint (4 weeks) Fine-tune a computer-vision or machine-learning model to flag exactly what matters in your data. The model runs on your own hardware, offline, with no per-call fees, and you own it along with the training data. Built with PyTorch, OpenCV, ONNX Runtime, and on-device inference. For teams with imagery or data that a general model does not understand well enough. ### AI Security Sprint (1 week) A one-week security audit run by two frontier AI models (Claude Opus and OpenAI Codex) against the OWASP Top 10, OWASP LLM Top 10, secrets, dependencies, infrastructure, and business logic. Every CRITICAL/HIGH finding is cross-verified by both models. You get a branded, board-ready report (audit register, domain scorecards, finding cards, prioritized rollout plan), Claude Code and Codex configured for your repo, and first-round fixes opened as pull requests. If the audit finds issues, the report is yours to keep no matter what you do next, and anything larger comes back as a fixed-price remediation sprint scoped straight from the findings. For regulated codebases (HIPAA, SOC 2, PCI), teams adopting AI coding tools, and founders facing an enterprise security questionnaire. ## Forward Deployed Engineer (FDE) Fractional or contractor Forward Deployed Engineer for companies that need AI deployed inside their business. Brad embeds with your team to scope, build, ship, and stay on the account until a real business metric moves. A fractional CTO retainer is available for ongoing work. ## Partners White-label AI delivery for software consultancies, audit and compliance firms, fractional CTO practices, and no-code agencies. Productized sprints (AI security audits, RAG, MCP servers, MVP builds) delivered under the partner firm's brand. ## Case study: Fitly AI Fitly is an AI-native fitness platform where one conversational agent does the job of five apps: it logs meals, builds and adjusts workouts, and tracks macros through natural conversation. It runs on 52 tools behind a ten-layer orchestration engine, where several systems run before the model is allowed to respond. Built with Claude Code and Codex from day one. https://www.pineaulabs.com/case-studies/fitly ## Example builds (scoped for clients) These are real engagements Brad scoped and quoted for clients, generalized here with client specifics kept private. - HIPAA-safe AI assistant (maps to the AI Integration Sprint): staff ask about any record in plain English and get instant, cited answers, with PHI de-identified before the model ever sees it and re-hydrated on display. https://www.pineaulabs.com/examples/ai-assistant - Computer-vision defect detector (maps to the Custom AI Model Sprint): a fine-tuned model that flags anomalies in imagery, wrapped in an offline tool the client owns, trained on licensed data. https://www.pineaulabs.com/examples/vision-defect-detection ## About Brad Pineau 25-year software engineer and former CTO. UPEI Computer Science graduate. Former CTO of the Timeless Group (Timeless Medical Systems), with 15+ years building HIPAA-grade clinical software where a data leak is a reportable event. Inventor on 4 patents. Now runs Pineau Labs, building fixed-price software and AI with agentic tooling. Based in Charlottetown, Prince Edward Island, Canada. ## Stack - Agentic AI tooling: Claude Code, Codex CLI, Claude API, OpenAI API, Google Gemini API. - Languages and frameworks: TypeScript, Angular, Next.js, React Native, Node. - Data and infrastructure: Postgres, pgvector, Pinecone. - Computer vision and ML: PyTorch, OpenCV, ONNX Runtime, model fine-tuning, on-device inference. - Platforms: web, iOS, Android, desktop. - Domains: HIPAA and healthcare IT (HL7, DICOM), fintech and payments (ACH), security auditing. ## Contact Email brad@pineaulabs.com, or use the intake form at https://www.pineaulabs.com/start to get a fixed-price quote. Replies come directly from Brad, over email by default, with Zoom available if you prefer.