Skip to content
View dylanpulver's full-sized avatar
3 followers · 0 following

Achievements

Achievement: Pull Sharkx4Achievement: Pair Extraordinairex4Achievement: YOLOAchievement: Quickdraw

Achievements

Achievement: Pull Sharkx4Achievement: Pair Extraordinairex4Achievement: YOLOAchievement: Quickdraw

Block or report dylanpulver

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
dylanpulver/README.md

Hey, I'm Dylan 👋

Canadian-American software entrepreneur and AI engineer based in Toronto. I help businesses use AI to run leaner.

  • 🔭 Current focus: agentic AI systems — AI that takes actions and completes work, not just answers questions
  • 🏗 A decade of fintech and automation platforms: crypto-exchange infrastructure, payments & KYC, RPA-driven bookkeeping automation, edge AI inference
  • 💳 Co-founder of GoldVault, a B2B payments platform
  • 🎓 Applied Mathematics @ Johns Hopkins · Master's in Mathematical Finance @ University of Toronto
  • 🎸 Also a musician and producer — the Dylan Pulver on Spotify and Apple Music is me

What I'm building

Project What it is
keypop Stop pasting API keys into AI agents — a native dialog collects secrets your agent never sees
session-deck Mission control for parallel Claude Code sessions — your agents, on one screen
judge-check Is your LLM judge discriminating or just nodding? Decoy-gated calibration for AI evals
crm-master A behavioral relationship engine — a personal CRM that does the discipline for you
prompt-creole A grammar for the language forming between humans and LLMs — documented from field evidence
code-factory A quality-enforcement factory for AI-written code — verify ladder, reviewer fleet, ratchet
research-factory Structured research that returns decision-ready, sourced briefs — and remembers them
ambient-wisdom Situation-triggered thinking tools for Claude Code, with a loop that learns what actually helps
prior-art Claude Code skill: scan the OSS/research landscape for projects doing what your repo does
agent-ops How to run Claude Code as a team of agents — session board, handoffs, state ledgers, overnight queue
claude-skills 20 battle-tested Claude Code skills in 5 installable plugins — the ones I run every day
recon Explainable transaction reconciliation — every match carries a receipt, property-tested
supabase-keepalive Keep free-tier Supabase projects from pausing — one GitHub Action, unlimited projects

I encode business processes into reusable AI skills — anything you'd train an employee to do can now be done your specific way, every time.

Find me

Pinned Loading

  1. keypop keypop Public

    🔑 Stop pasting API keys into Claude Code & Cursor. keypop pops a native dialog to collect a secret your agent never sees — value goes to a local file, never the model's context. 60 lines of shell, …

    Shell

  2. session-deck session-deck Public

    Mission control for parallel Claude Code sessions — a zero-dependency live dashboard over the agent-ops session board. Your agents, on one screen.

    JavaScript

  3. crm-master crm-master Public

    A behavioral relationship engine — a personal CRM that does the discipline for you. Nudge queue, warmth decay, AI drafts that learn from your edits.

    TypeScript

  4. claude-skills claude-skills Public

    Battle-tested Claude Code skills I run every day — 20 skills in 5 installable plugins: session continuity, shipping, thinking tools, launch, and skills governance

    JavaScript

  5. recon recon Public

    Explainable transaction reconciliation — a five-tier matching ladder where every match carries a receipt: which rule, which fields, how much tolerance it consumed. Integer money, property-tested in…

    TypeScript

  6. agent-ops agent-ops Public

    How to run Claude Code as a team of agents — session board, handoffs, state ledgers, config sync, and an overnight queue. Patterns run daily across 3+ parallel accounts.

    Shell