Skip to content

CV

Read it here, or take a copy.

Download PDFEmail me

Mayowa Francis Dada

AI Product Engineer

United Kingdom · mayowa.dada@solat.io

linkedin.com/in/mayowadada · github.com/Haryohmi

Profile

AI Product Engineer who takes products from an ambiguous problem to running software: architecture, backend implementation, deployment, and the operational work that follows. I build with large language models, structured LLM workflows and production Python services, and I am as interested in what a system cannot do as what it can. I currently own the engineering for a live AI product with paying users, and run SOLAT Technologies delivering client software.

Core technical skills

Languages
Python · TypeScript · SQL
Backend
FastAPI · REST APIs · PostgreSQL · Redis · Async processing
AI systems
LLM application design · Structured prompting · Long-context pipelines · LangChain · Google Gemini · Prompt evaluation
Web
Next.js · React · Tailwind CSS · Payload CMS
Production
Vercel · Supabase · Stripe · Resend · Sentry · PostHog · Chrome extensions (MV3)
Quality
Vitest · Playwright · axe-core accessibility testing · CI pipelines

Experience

Founder & AI Product Engineer — SOLAT Technologies

Jun 2025 – Present

  • Built and operate TubeMate, a live AI YouTube learning assistant on the web and as a Chrome extension. Python/FastAPI, PostgreSQL and Redis, with Gemini 2.5 Flash Lite via LangChain. 224 registered users, 130 summaries across 94 videos and ~79 hours of content processed as of August 2026.
  • Chose long-context prompting over retrieval-augmented generation after measuring that target transcripts fit inside the model's context window — removing a vector index, its latency and a class of retrieval failure, and tying the product's video-length tiers to that constraint.
  • Diagnosed and fixed a production incident where YouTube rate-limiting triggered a retry cascade (~70 Sentry events), adding a Redis-backed cooldown to break the loop.
  • Reduced registration abuse with a layered response — an email verification gate on product use, per-IP registration and login limits, and fingerprint-based throttling on the unauthenticated endpoint.
  • Shifted the product's success measure from raw signups to verified and activated users after finding that nearly half of registrations never verified an email address.
  • Integrated Stripe billing, transactional email, error tracking and product analytics.

Full-stack Engineer — client delivery — via SOLAT Technologies

2025 – Present

  • Community platform (Next.js, Payload CMS, PostgreSQL). Led information architecture and build. Chose Payload over a hosted CMS because a nine-role permission model, audit logging and submission-queue ownership exceeded what hosted tooling offered without enterprise pricing plus a separate admin application. Modelled 14 content collections with distinct access rules and publishing lifecycles. WCAG 2.2 AA checks run in CI alongside typecheck, lint and migration-backed builds.
  • Booking and payments platform (Next.js, Supabase, Stripe). Built the customer booking flow, an availability engine covering eight validity rules with DST-correct London-to-UTC handling, an admin dashboard and calendar sync. Used Stripe's hosted Checkout so card details stay within Stripe's interface. Unit-tested the pricing and availability logic, where double-bookings and payment edge cases carry direct commercial cost.
  • Made-to-order retail storefront (Next.js, static). Identified that a self-service cart did not match a business selling through conversation, and built a collection-led catalogue routing interest to direct enquiry instead — a scope decision visible in the codebase, which carries no cart, checkout or inventory layer.

Software Engineer (contract) — UK startup

Jun 2024 – Sep 2024

  • Developed backend services and REST APIs in Python and Node.js, including document-processing and AI-assisted CV generation workflows.
  • Contributed to frontend development and architectural decisions on scalability and maintainability.

Selected projects

TubeMate — live AI product, web and Chrome extension. Freemium with Stripe billing.

ÒreAyò — active research into speech recognition for Nigerian languages (Yoruba, Igbo, Hausa). Pre-release.

Education

MSc Management of Business Information Technology — Distinction

BSc Psychology

Additional

Alongside my technology work I am employed in a regulated custodial environment in the UK, which has developed my approach to responsibility, documentation and decision-making under pressure.