█████╗  ██████╗ ███████╗███╗   ██╗████████╗███████╗              
██╔══██╗██╔════╝ ██╔════╝████╗  ██║╚══██╔══╝██╔════╝              
███████║██║  ███╗█████╗  ██╔██╗ ██║   ██║   ███████╗              
██╔══██║██║   ██║██╔══╝  ██║╚██╗██║   ██║   ╚════██║              
██║  ██║╚██████╔╝███████╗██║ ╚████║   ██║   ███████║              
╚═╝  ╚═╝ ╚═════╝ ╚══════╝╚═╝  ╚═══╝   ╚═╝   ╚══════╝

> AI R&D and Product Development

We work on the AI problems that don't have a playbook yet.

We build experimental, frontier AI systems and ship them to production — from zero-to-one greenfield builds to AI woven into platforms that already serve millions. We work directly with founders and executives.

> We run what we've built in production — or train your team to.

Our Services

✻ Research & Prototyping

Explore what's possible before committing to a build. We research, prototype, and validate AI approaches so you invest in what actually works.

▶ Product Development

Zero-to-one builds or integration into existing systems. We don't build demos — everything we ship is on a path to production.

◉ Models & Infrastructure

Frontier APIs, open-weight self-hosted models, or classical ML — we work across the full stack and bridge traditional AI/ML with genAI and LLMs.

⚑ VC Services

Portfolio intelligence platforms, fund accounting systems, and deal flow engines built for the largest VC and PE firms.

Work With Us

> We work best with Seed–Series B founders who are also the operator — and want a hands-on partner committed to the outcome, not a vendor at arm's length.

Leadership · Strategy · Retainer

◈ Fractional CTO & CAIO

An experienced operator embedded with your leadership — owning AI and engineering strategy, and shipping it.

Who it's for

Seed–Series B founders who are also the operator and want a long-term partner, not a strategy deck.

What you get

  • Monthly strategy and operating cadence with your leadership team
  • An AI and engineering roadmap grounded in what to build, buy, or wait on
  • Hands-on architecture, hiring, and code review across your stack
  • We run what we recommend in production — or train your team to

Agents · Infrastructure · Deployment

◆ AI Agent Software Factory Implementation

Stand up a closed-loop agentic software factory — humans and coding agents shipping production code through one pipeline.

Who it's for

Technical founders and teams who want agent infrastructure in production without losing engineers to weeks of setup.

What you get

  • A multi-agent coding pipeline — Claude Code, Codex, Cursor, autonomous loops — on one closed loop
  • Multi-worktree orchestration, cross-model review with disagreement gating, branch-per-PR databases
  • A compound workflow where every bug becomes a regression test and every fix a reusable skill
  • Documented setup your team owns outright, plus a hypercare window after handoff

Background

Here Be Dragons is led by Dan Andrews — a startup CTO, venture-backed founder, and startup studio founder with successful exits in the AI space. Alongside HBD, he advises a portfolio of early-stage AI startups backed by Y Combinator and Techstars.

Startup CTOVenture-Backed FounderStartup Studio FounderSuccessful ExitsYC & Techstars Advisor

Featured Work

> What we've shipped across healthcare, fintech, AI infrastructure, venture capital, and sales enablement.

◆ Self-Improving Agentic Software Factory

Coding Agents · Compound Engineering · Multi-Agent Orchestration

Designed and recently shipped a Level 7-8 software factory for a venture-backed startup, where humans and coding agents (Claude Code, Codex, Cursor, GitHub-native, autonomous loops) deliver production code through one closed-loop pipeline. Every bug becomes a regression test, every incident a runbook, every repeated agent mistake an updated AGENTS.md, skill, or eval. Multi-worktree parallel orchestration, cross-model review with disagreement gating, scoped agent identities, branch-per-PR databases, and a post-merge compound workflow that turns each shipped change into reusable skills and playbooks.

◆ IRB-Approved Voice AI for Clinical Trials

Healthcare · Voice AI · Regulatory Compliance

Built among the first AI voice agents approved by two of the world's largest IRBs to handle clinical trial prescreening, informed consent, and participant scheduling. Voice-native agents (not chatbots) embedded in the trial workflow: prescreening qualifies inbound candidates and routes them forward; informed consent walks enrolled participants through the consent process by voice, answers questions in real time, and checks comprehension; scheduling handles visit booking, reminders, and changes. All three escalate to humans for clinical questions and live as platform-native workflows rather than bolted-on widgets.

◆ Prediction Market Intelligence & Trading

Classical ML · Multi-Agent Research · Autonomous Trading

End-to-end prediction market platform that aggregates data across five major exchanges, detects cross-platform arbitrage via fuzzy matching, and executes trades using Kelly Criterion sizing. A multi-stage AI research pipeline (Gemini for research, Claude for synthesis) feeds probabilistic confidence scores into classical ML trading models — bridging traditional quantitative methods with LLM-driven analysis.

◆ Contact Intelligence & Sales Enablement Platform

Deep Research Agents · Multi-Model Routing · VC-Backed Startup

Built the AI engine for a venture-backed sales enablement platform. 30+ specialized extraction agents for contact enrichment, resume parsing, risk scoring, and fake profile detection. Multi-model routing — Gemini Flash for fast inference, Claude for complex reasoning — across a production API serving millions of enrichment requests.

◆ Academic Credential Verification at Scale

Deep Research Agents · ORCID · Scholar APIs

Built specialized deep research agents for the largest data labeling platform to verify PhD credentials and research publications. Cross-references ORCID, Google Scholar, Semantic Scholar, CrossRef, and university databases to validate academic claims — conference presentations, publication records, field-of-study matching, and institutional affiliations.

◆ Fully Agentic Back Office

SaaS Operations · Autonomous Agents · YC-Backed Company

Designed and deployed a fully autonomous back office for a mobile application marketplace — handling customer support tickets, inbound queries, user account management, subscription billing, refunds, and FAQ resolution. Zero human operators in the loop for standard operations.

◆ Post-Training Evaluation for Coding LLMs

LLM Infrastructure · Eval Harnesses · A/B Testing

Built automated post-training evaluation loops and test harnesses for agentic coding models at one of the largest LLM providers. Automated A/B testing of coding models against real-world programming tasks at scale, providing rapid feedback on model quality between training runs.

◆ Portfolio Intelligence for VC & PE Firms

RAG · Fund Operations · Hybrid Search

Built portfolio intelligence platforms for several of the largest venture capital and private equity firms. Reverse portfolio search engines, automated fund accounting workflows, and large-scale data ingestion pipelines transforming unstructured deal flow into structured, searchable knowledge bases.

◆ VC Due Diligence & Deal Flow Automation

Workflow Automation · Research Agents · Accelerator Programs

Built deal flow research and due diligence automation for multiple well-known accelerator and VC programs. Automated intake triage, founder background research, market sizing, and competitive landscape analysis — turning weeks of manual due diligence into hours.

◆ Autonomous AI Video Production

Content Generation · Video Rendering · Multi-Platform Publishing

Built a fully autonomous video production pipeline — AI-generated scripts, local TTS, AI-generated B-roll, animated data visualizations, subtitle burning, and section composition via FFmpeg. Publishes finished videos across TikTok, X, and other platforms. End-to-end from research to published content with no human in the rendering loop.

Ideas in Motion

> Alongside client work, we build our own tools and concepts. Here's one in progress.

✦ Rekko

Building

Prediction Markets · Trading Signals · API & MCP

An intelligence layer for prediction markets — monitoring 1,000+ markets across Kalshi, Polymarket, Robinhood, and Coinbase. Rekko detects cross-platform arbitrage and turns live order flow into execution-ready trading signals, delivered over a REST API and MCP server built for traders, bots, and AI agents.
rekko.ai ↗

Core Capabilities

Custom MCP Servers
Agent Skills & Orchestration
RAG Solutions
Prompt & Context Engineering
Vector Databases
Semantic & Hybrid Search

About Us

What We Do

Old maps marked uncharted territory with “here be dragons.” We took the name because that's where we work — on the AI projects that don't have a playbook yet. Experimental systems, novel architectures, problems that require real R&D.

Founded in 2009, we work directly with founders and executives at venture-backed startups and large enterprises. Every project ships on a path to production — whether that's a greenfield build from zero or integrating AI into systems that already serve millions.

◇ Our Mission

Take experimental AI from research to production. We work on the projects that are too new for a vendor solution and too important to get wrong.

↯ Our Approach

Embedded with your leadership. Fast cycles from research to prototype to production. We work across the full model stack — frontier APIs to self-hosted open-weight models.

Team

DA

Dan Andrews

Founder & CEO

15+ years building products at startups and enterprises. Bridges the gap between AI research and production systems.

Contact

Let's Talk

We take on a small number of engagements each quarter. If you're building something interesting in AI, we want to hear about it.