Software Engineering
- Critical path shipped
- Production seams fixed
I will audit, review, and fix your vibecoded shitshow.
Highly weaponized, military-grade autism solving your engineering challenges.
You hire me for the ugly systems work: sit with the broken thing, spot contradictions, delete noise, and make the code safe enough for users.
01
Repo, branch, bug, outcome. I start at the contradiction.
02
I trace the pattern across state, data, prompts, and deploys.
03
Commits, tests, notes, and code the team can explain.
The app works until reality touches it: auth, users, payments, weird state, and the second click.
AI got you a demo. My autistic focus earns its keep after the dopamine, where I stay with contradictions until I can name why the code fails.
Pick the engagement that matches the failure path. Audit, repair, training, and technical leadership all end in written decisions and working code.
You pay for obsessive attention to the part the team keeps talking around, then the discipline to make it boring.
Production damage, shipped systems, and code that had to keep running under real constraints.
Rust services, React migrations, realtime pipelines, AI runtimes, latency work, and the details my brain refuses to ignore.
You have a repo full of plausible answers and no proof the product can survive users. I bring autistic pattern recognition, senior engineering judgment, and the patience to chase the weird path until the fix is clear.
No vague retainer fog. You get the autistic inspection first: what breaks, what matters, and which repair buys the most safety.
I have shipped software for 19 years and spent the last 2 fixing vibecoded SaaS messes for founders. The useful part is the same: pattern recognition, low tolerance for sloppy abstractions, and enough focus to finish the repair.
blprnt
Built a local AI execution runtime for autonomous teams: Rust coordinator, durable state, multi-provider model routing, and a React/TypeScript operating surface.
Amazon Web Services
Led frontend modernization inside AWS, moving a legacy Java UI toward React and TypeScript with tighter delivery, accessibility, and review standards.
Trading Systems & Consulting
Delivered low-latency Rust services, real-time ingestion pipelines, and data-dense React workflows across fintech, ecommerce, and SaaS products with hard failure boundaries.
The bot drafts from context. I check the parts it treats as disposable: four owners for one value, auth that lies in one route, and tests that missed the bug.
| Decision point | Hire An Autist | AI-only pass | Typical agency |
|---|---|---|---|
| First useful output | A blunt map of what is broken and where to cut | More code on top of the same confusion | Discovery before someone opens the repo |
| Debugging style | Autistic pattern lock across state, data, prompts, and deploys | Local guesses from whatever context fits | Tickets routed through layers |
| Best fit | Ugly systems, AI slop, brittle UI, strange backend behavior | Fresh demos and shallow fixes | Greenfield builds with room for meetings |
| Communication | Specific written tradeoffs with no softened risk language | Confident paragraphs | Decks, standups, and consensus fog |
The questions founders ask after they send the repo and before they decide whether the blunt branding is a bit.
Within a week in most cases. Send a clean brief and I can move sooner. A vague calendar for next quarter belongs with someone else.
Then I say so. The damage report gives you the least stupid next move: fix, rebuild, cut scope, or bury it.
Yes. Short audits, fractional advisory, and fix-and-ship blocks work. The scope follows the failure path.
Yes. I teach teams where AI helps, where it lies, and how to review generated code before it mutates the product.
The mess decides. Audits stay contained. Last-mile fixes scale with risk, ownership, and how much generated code we need to unwind. I quote numbers, not vibes.
Diagnosed. My autism shows up in the work: pattern lock, low tolerance for fuzzy logic, and long focus runs on problems other people avoid.




Send the repo, the broken flow, or the AI-generated knot the team keeps avoiding. I will read it, name the failure path, and tell you where my attention helps most.