BASED IN
Toronto, Canada
I scope, build, and ship production-grade AI workflows on a code-first stack — Claude Code, Vercel cron, serverless. No Zapier sprawl, no $300/mo platform fees. Just the workflow you needed, deployed and observable, in days not quarters.
Productized so you know what you're getting and when. Fixed scope, fixed delivery, no retainer creep. Pricing is call-first — every project has different signal. See full service descriptions →
A week sitting with your business: founder interviews, workflow mapping, finding the 3-5 places AI gives back real time. Walk away with a prioritized build plan, vendor recs, cost projections, and a stack-of-cards diagram.
Productized scope. We agree Monday morning, I deliver deployed + documented by Wednesday end-of-day. If it's not "send this to my inbox at 6am" specific, the sprint doesn't fit — and I'll tell you that on the call.
For founders already building with AI who need a second brain. 2x/month video, async DMs between, prompt + cost + architecture reviews. Not implementation — the kind of feedback that saves you from shipping the wrong thing.
Multi-workflow builds that need more than a sprint. Content engines, competitor watch, Whop onboarding, MCP servers, full marketing sites with SEO infrastructure. We scope it on the call, I quote it written, you decide.
Three from clients · names redacted on request until case studies publish in Q3 2026.
The audit alone paid for itself in week two. We deferred two SaaS subscriptions because Pravine's report showed we'd been Zapier-stitching a problem that wanted 40 lines of code.
I've hired three "AI consultants" before. Pravine is the first one who shipped working code instead of a Notion doc. Build sprint hit the deadline to the hour.
Async-first actually works the way he runs it. Daily Looms, written summaries, no meeting tax. The CISNET migration ran 3 months without a single live call drift.
What surprised me: he turned down two scopes that weren't a fit. Refunded a sprint that hit a dead-end. That's worth more than the work.
No "discovery deck," no 60-min kickoff calls just to align timezones. Lean, written-first, async-defaulted. Full breakdown at /services.
30 min · within 48 hrs. 4 questions on the booking form: what's broken, who you are, when you need it, rough budget. I read everything personally.
Operator-to-operator call. Sharpen the scope, agree on service tier, identify fit-killers early. No pitch deck. If it isn't a fit, I'll point you to who is.
Same day. Fixed price, payment terms, milestones. You sign with a Stripe link. Work starts when you say go — usually within a week of signing.
Daily progress notes. Milestone demos via Loom. Final handover with README, deployment guide, and a Loom walkthrough you can revisit.
Two groups, split by status. Live links where applicable. The full picture — including screenshots, stack chips, and per-project metrics — lives at /work.
Opinionated, not exhaustive. Tools I actually run · daily-rotation marked green. The full landscape map lives at /community/landscape. Full stack notes at /portfolio/stack.
Toronto-based solo operator. Started shipping web apps in 2022; pivoted hard into AI ops in 2024 when Claude Code dropped and "automate this in a weekend" became actually true.
Two things, full-time: shipping AIOpsForge (a $9.99/mo community teaching solo founders to build with AI) and freelance AI automation for indie founders + small agencies.
Fast because I have a strong opinion about the stack. Claude Code + Vercel + a few well-chosen SaaS = ~80% of what most founders need. The other 20% is the work. More on the about page →
Common ones up top. Detailed answers (and 6 more) live on /services.
30-minute call · zero pitch · you describe the problem, I tell you whether AI is the right fix and if so which service tier matches. If it isn't, I'll tell you that too and point you to who is.