Chadi Choukkah
Software engineer: TypeScript, React and Postgres, with test automation.
Agadir, Morocco · open to QA, full-stack and front-end roles, remote or relocating
caught by Rihla's tests. Tap any figure to see how it was measured.
·
Who's reading?
Pick one and the tour leads with what matters to you.
Everything, in the order the tour takes it
·
I test software, and I build it.
Building software since 2019; these days mostly test automation and full-stack TypeScript with Next.js. I've spent two years applying other people's guidelines on model evaluation. On my current projects I write the specifications and the rules, AI coding agents do much of the implementation, and I review, test and decide what ships (). One is a live site with . The other is a freight platform with that isn't deployed yet. Between them they have .
Figures about the projects today are recounted from their source code every Monday, most recently on 16 November 2026. Figures inside the stories stay as they happened. Open any of them to see how it was measured.
Sightline live
A site that tells you where a match, film or series can legally be watched from wherever you are. Rights get sold country by country and expire on dates nobody announces, so the real problem isn't fetching data. It's stopping a correct answer from going wrong on its own.
A personal project, built by directing AI coding agents: I set the specification and the checks, and review what ships.
1,555 unit and component tests across 128 files, run again by this site every Monday against the pushed code. Provider failures are injected at the network layer so the real code path runs. Both rate limits were proven under concurrency on localhost: 180 requests, exactly 120 served and 60 rejected.
Test types unit · component render ·
fault injection · contract (provider response shape) · structural (rules asserted over the
source tree) · static type assertion · end-to-end
Tools Vitest · Playwright across Desktop Chrome, Mobile Safari and Mobile Chrome ·
Testing Library with happy-dom · MSW · Zod · @ts-expect-error compiler gates ·
tsc --noEmit
Failure modes exercised timeout · HTTP 500 · HTTP 429 · malformed JSON ·
provider down · empty result vs faulted lookup
Every call to an outside provider goes through a Redis cache, and a test fails if one doesn't. The cache keeps stale data physically present so it can be served when revalidation fails, an outbound quota per provider is shared across serverless instances rather than kept per process, and provider responses are parsed by Zod schemas, 8 schema files so far.
Resilience patterns
stale-while-revalidate with a 48-hour grace window carrying the answer's age ·
negative caching of 404s · in-flight request deduplication · exponential backoff
(200/600/1800 ms) · fault memoisation so one caller discovers an outage and the rest are
told without a network wait · provider fallback with strict precedence, never a merge
Rate limiting inbound fixed-window per caller (Redis INCR with a
conditional EXPIRE) · a separate per-provider outbound budget, because the two
solve opposite problems and were once confused in a real defect
Validation Zod schemas at provider boundaries, with types inferred from the schema
so a nullable field becomes a compile error rather than a runtime 500
Stack Next.js 15 App Router · React 18 · TypeScript · Upstash Redis · Vercel ·
GitHub Actions · no database and no user accounts
Integrations ESPN · TheSportsDB · TMDB · Wikimedia Commons and Wikidata
HTTP and data standards 429 with Retry-After ·
Cache-Control: no-store on degraded responses · 404 as absence and 5xx as
fault · 22 Schema.org types · ISO 3166-1 alpha-2 · ISO 8601 · IANA time zones ·
sitemap protocol · Web App Manifest
46 page routes and 183 components. Player and fighter photos appear only when Wikimedia Commons confirms the licence, with the credit generated from the same answer. The site also carries a FIFA World Cup 2026 archive of 104 matches, built from ESPN data. Analytics load only after consent, with accept and reject given identical visual weight.
Rendering React Server Components ·
streaming with Suspense · static generation via generateStaticParams ·
time-based revalidation · explicit dynamic opt-out where freshness matters
Core Web Vitals, by metric targeted fetchPriority="high" on the LCP
image · icons generated at build time from one SVG, 611 kB to 9 kB · a server-side slice
that cut 381 images and 2,322 DOM nodes down to what the page had actually asked for
Accessibility, by success criterion 4.1.2 Name, Role, Value · 2.1.1 Keyboard ·
2.4.3 Focus Order — combobox roles, keyboard navigation and Escape restoring focus, with
tests · 1.1.1 Non-text Content via aria-hidden on decorative slots
Security headers Content-Security-Policy with a per-response nonce and
strict-dynamic · HSTS with preload · X-Frame-Options ·
X-Content-Type-Options · Referrer-Policy ·
Permissions-Policy disabling camera, microphone and geolocation
Privacy consent-gated analytics with accept and reject given identical weight
Screenshot taken automatically from the live site
Security, mapped to OWASP categories
A03 Injection — JSON-LD escaping of <, with tests, plus a nonce-based CSP ·
A03 — schema validation at provider boundaries · A04 Insecure Design — rate limiting ·
A05 Misconfiguration — security headers and a repo-hygiene test that fails if an
ignored file is tracked · A10 SSRF — an allowlist of hosts on the image proxy
Rihla not yet deployed
An offline-first app for a carrier moving baggage and freight between France and Morocco, designed with a real carrier to replace paper registers and Excel; its staff began testing it hands-on in October 2026. The interesting problem is authorisation: the same person is often the owner and the driver, so the system has no roles at all. Permissions are capabilities, and Postgres enforces them at the row level rather than trusting the application to remember.
A personal project, built by directing AI coding agents: I set the specification and the checks, and review what ships.
1,398 unit tests, and a nightly run that breaks each guard on purpose in a throwaway copy to prove it still fails. I added mutation testing with Stryker on the business rules: 1,468 deliberate changes to the code, with the score raised from 83.99% to 98.30%, closing edge cases that let a driver deliver the same parcel twice or see another driver's cash movements. I checked the tool too: Stryker reported 25 surviving mutants, and applying each by hand showed 17 were already killed. When I automated the policy count for this page, the script said 51 and my hand count had said 52. Both were wrong: one migration generates 17 more in a loop at run time, so the real number was 68. The counter now replays the migrations instead of matching text, and refuses to answer if it meets a statement it can't expand.
Test types unit · integration · mutation
testing with a measured score · guard self-tests, each guard broken nightly to prove it still
fails · static analysis · policy tests on real Postgres inside transactions that always roll
back (6 of the 68 policies so far) · end-to-end on five device profiles (signed-out screens
so far)
Tools Vitest · StrykerJS · Playwright · dependency-cruiser layer rules · two custom
ESLint rules, for right-to-left layout and for numbers in Arabic text · lefthook
21 tables, 60 foreign keys, 68 row-level security policies, 8 parameterised capabilities instead of roles. Writes are queued in the browser first and drained later, so the same record can exist on a phone that has been offline for hours and on a server that has moved on.
Authorisation capability-based access
control rather than RBAC · PostgreSQL row-level security with
FORCE ROW LEVEL SECURITY · SECURITY DEFINER functions · JWT
custom access token hook · multi-tenancy with fail-closed policies
Offline outbox pattern · IndexedDB via Dexie · queue drain with idempotency keys ·
device-precedence conflict resolution (no trustworthy clock)
Stack TypeScript monorepo of six layers, with import direction checked on every commit ·
React 19 · Vite · Supabase Postgres and Auth · PostgREST
Domain standards ISO 3166-1 · IANA time zones · multi-currency with no implicit
conversion
20 screens, offline-first, with the interface in French and Arabic — real right-to-left handling rather than a mirrored stylesheet.
Internationalisation French and Arabic ·
RTL with dir switching · logical CSS properties · isolated Latin digits and
codes
Offline UX a service worker, so the app opens with no signal · optimistic writes ·
queued actions with visible state
Stack React · Vite · Tailwind with logical properties only · IndexedDB
Sports live scores and data platform production
Built for an agency client, to their requirements and their review, from November 2025 to July 2026; I maintain it on request. A Next.js 14 platform with 43 pages and 44 API route handlers, putting four sports and news sources (TheSportsDB, ESPN, NewsData.io and UFC.com) behind one validated data layer.
A tiered cache, Upstash Redis with an in-process fallback, with TTLs set by how fast each kind of data changes: thirty seconds for live scores, up to twenty-four hours for team and league data. I also traced a site-wide indexation failure to a canonical-domain misconfiguration and automated search submission through IndexNow and the Google Indexing API in CI.
Core Web Vitals work across a large dynamic route surface, SCSS and Tailwind, built for a client rather than for myself.
Stack Next.js 14 App Router ·
TypeScript · Redis · REST integrations across multiple sports data providers · Vercel
Patterns incremental static regeneration · dynamic route generation ·
server-side caching
Front-end React · Tailwind · SCSS · Core Web Vitals · responsive dynamic routes
AI and LLM model evaluation 2024–2026
Contract work with Outlier AI and UseBrainPlatform: evaluating model outputs and reviewing chain-of-thought reasoning, with Python tooling around the review work.
The work is adversarial reading — following a chain of reasoning to find where it becomes confidently wrong rather than where it stops. It is the same instinct the rest of my work runs on, which is why these projects are mostly about telling a real answer apart from a plausible one.
Evaluating model output is testing without a spec: looking for the case that passes every surface check and is still wrong.
Work model output evaluation ·
chain-of-thought review · comparative response assessment · annotation against task
guidelines
Platforms Outlier AI · UseBrainPlatform
Tooling Python
Data automation — HCI AirCom, China (remote) 2019–2020
A Python and Selenium scraper that cut a 900-company data run from thirty minutes to under five, replacing a manual collection process for a team that used the output.
Where can I watch this, from where I am?
Broadcast rights get sold country by country, and they expire on dates nobody announces. Fetching the data is the easy half. The hard part is stopping a correct answer from going wrong on its own.
Live at sightlinetv.com · Next.js on Vercel · most of the code written by AI coding agents at my direction, so "I" below means what I decided, directed or checked
What this holds, and what it doesn't. Started on 11 August 2026 from an earlier sports-data codebase, it holds across 7 countries and 4 competitions, checked by hand. No feed updates them, and every one can go stale without anything telling me, which is the problem the rest of this page is about.
The one idea it's built on
Failing to fetch data isn't the same as data telling you there's nothing there. In code both arrive as an empty array. To a reader they mean opposite things. Mix them up and the site starts answering confidently when it should be saying it doesn't know, and that's worse than being down.
Publishing a broadcaster the user cannot actually watch on is worse than publishing nothing — it is the one claim this whole site is built on. lib/data/broadcast-rights.ts:11
Follow a thread
Nobody sells this data. I looked. The feeds that exist carry fixtures and metadata, and you can infer rights from them if you want to, but a guess printed as a fact sends someone to a channel they don't have.
So the table is typed by hand. It grows at whatever speed I can check things, which is slow, and every season's turnover is more manual work. A feed would have covered thirty countries on day one. I'd rather have 7 I can stand behind.
Three weeks into the season, the Australian row said Optus Sport. Optus had handed its rights to Nine on 30 June and shut the service down completely on 1 August. So the site was telling Australians to subscribe to something that didn't exist any more, and had been doing so since the first weekend.
The mistake isn't the interesting part. "Verify these before deployment" had been sitting in my progress file since July, doing nothing.
A note in a queue is not a mechanism.lib/data/rights-freshness.ts
What replaced it: listings can carry an expiry date, and everything reads through a filter instead of the raw array. , so I can't honour the expiry in one place and forget it in another. A competition whose listings have all lapsed drops off the site rather than rendering an empty panel.
That same claim was also sitting in a blog post, which the expiry filter never touches.
So I fixed the post. I edited lib/blog/posts.ts, and the first line of that
file says it's auto-generated and not to be edited directly.
My local build runs Next straight and skips the generation step. The deploy doesn't. Typecheck passed, the tests passed, the build passed, the deploy went green, and production carried on serving Optus.
I lost a good while convinced it was cached. It wasn't. Cache miss, age zero, no-store. The edit had simply never survived.
Regenerating from the real source also cleaned two files I'd never opened. The claim had been living in four places. My mechanism reached one of them.
Faults get recorded separately from results. If a lookup failed, you see "we couldn't check just now". If it worked and found nothing, you see "nothing recorded". A country I've never asked about is a third thing again.
It matters more than it reads. If a lookup times out and the page 404s, Google takes that as permission to deindex, and getting back in takes far longer than the outage did.
A missing API key fell through to "123", which is TheSportsDB's shared
public test key. It answers. Valid JSON, a slice of the catalogue, so a missing key looked
like a working one at every layer above it.
There's a longer list. A rate limiter guarding routes I'd already deleted. A dependency audit reading a lockfile the project doesn't install from.
The one with a number on it: I deleted an environment variable nothing read. Nothing did read it. But middleware threw on it, and every route returned 500 for about .
None of these were missing checks. Every one had a check that passed, confirming something next to the thing I was trusting it for. So the assertions moved onto what the site actually serves instead of how it's configured.
No, and it took three tries to work that out. I quoted a Lighthouse score of 99. Four later runs of the same build gave me 92, 94, 93, 92.
So I built a proper harness. Pinned browser, fixed throttling, five runs, median and spread printed together. It told me the homepage was failing, I wrote that down as framework-bound and not worth chasing, and I reverted a change because the numbers didn't support keeping it.
The harness was the problem. Its own spread ran over thirty points, from CPU contention and no warm-up run, so it couldn't see an effect smaller than its own noise. That's exactly what I'd pointed it at. On Google's hardware: . I put the reverted fix back.
I'd built a tool to stop myself quoting single runs, then trusted it without asking whether it could see anything.
What it looks like
Screenshots taken automatically from the live site
About
I'm a software engineer in Agadir, Morocco. I've worked on contract and on my own projects, which means I've carried them end to end: schema design, backend, interface, test strategy, deployment.
Working alone doesn't mean unreviewed. The sports platform was built to an agency's requirements and reviewed by them. Rihla came out of sitting with an operator, reading his paper documents and asking what he actually does on a Tuesday. Two years of model evaluation is nothing but applying someone else's guidelines consistently and being marked on it. I'd like a team next, and I'm looking for full-stack, front-end or test automation roles, remote or relocating.
There's a thread running through the work. I don't like systems that go wrong quietly, so I build for the case where a provider is down, a deal has expired, or a check is passing for the wrong reason. Telling "no answer" apart from "no attempt" turns out to be most of the job.
How I build
I write the specification and the rules, AI coding agents (Claude Code) do much of the implementation, and I review, test and decide what ships. Agents are confident and often wrong, so each project carries written rules the agent reads before every session, and checks that can fail: a test only counts once I have seen it go red, and mutation testing shows whether the tests notice wrong code.
Two catches that shaped how I work. An agent writing documentation produced a code snippet and attributed it to a test that never contained it; since then every quoted string is checked against the file. And on Rihla, StrykerJS reported 43 mutants as surviving; checking them by hand showed 32 were already caught, so the tool's report is now verified before any test is written against it.
Experience
Rihla — software engineer, personal project
Offline-first app for freight carriers, in Arabic and French, designed from a real carrier's paper register and tested by its staff since October 2026. Postgres row-level security, mutation testing, React.
Sightline — software engineer, personal project
Live site answering where a match, film or series can legally be watched, country by country. Next.js, caching built around freshness, fault injection in tests.
Full-stack developer — sports live scores platform
Next.js 14 and TypeScript, four sports and news sources behind one validated data layer, Redis caching.
AI model evaluator — Outlier AI, UseBrainPlatform
Evaluating model outputs and reviewing chain-of-thought reasoning against detailed guidelines, grading each issue by severity.
Web developer — Wino Trans, Agadir
Built the company's first website with a live luggage-pricing calculator as a credited internship, then was rehired to rebuild the front end in SCSS with interactive pricing.
WordPress developer — Silyana Jewelry (France)
Built the designer's layouts as a working WordPress store site.
Business development specialist — HCI AirCom, China (remote)
Python and Selenium scraper that cut a 900-company data run from thirty minutes to under five; the contract was extended to more tasks.
Education
- B.Sc. Software Engineering — Üsküdar University, Istanbul
- Sep 2019 – Jul 2023 · taught in English, 242 ECTS · High Honor (Spring 2020), Honor (Spring and Fall 2021) · two credited internships: an Android voting app for university elections (Aug–Sep 2021) and the Wino Trans website (Aug–Sep 2022)
What I work with
- Languages
- TypeScript · JavaScript · Python · SQL, including Postgres row-level security policies · Java
- Front-end
- React · Next.js App Router · Vite · Tailwind · SCSS · Core Web Vitals · keyboard and screen-reader accessibility · internationalisation including RTL
- Back-end and data
- Node · PostgreSQL · Supabase and PostgREST · row-level security · Redis (Upstash) · REST integration · caching strategies including stale-while-revalidate and negative caching · rate limiting (per-IP windows, token-bucket quotas)
- Testing
- Vitest · Playwright · Selenium WebDriver · JUnit · Postman · API testing: contract tests against recorded provider responses and network-level fault injection with MSW · unit, component-render, fault-injection and structural testing · static type assertions as compile-time gates · mutation testing with StrykerJS · coverage gates enforced in CI
- Infrastructure
- Vercel · GitHub Actions · Git · scheduled jobs via cron workflows · dependency auditing as a pipeline stage · Content-Security-Policy with a per-response nonce
- Standards I work against
- WCAG 2.2 success criteria 4.1.2, 2.1.1, 2.4.3 and 1.1.1 · Core Web Vitals · OWASP Top 10 categories A03, A04, A05, A06, A10 · Schema.org · ISO 3166-1 · ISO 8601 · IANA time zones · GDPR consent gating
Languages
- Tamazight and Arabic
- Native
- English
- C1
- French
- B1–B2
- Turkish
- Conversational
Contact
forchadiwork@gmail.com
linkedin.com/in/chadi-choukkah-802066177
github.com/shad-yy
Try one
Three problems from building Sightline. Pick what you would do, then see what I did, including the turns that were wrong.
Nobody sells the data you need
A site has to say where a match can legally be watched, country by country. Rights are sold per country and expire on dates nobody announces. What do you build on?
The fix shipped. Production still serves the old answer.
You edited the file that held a wrong claim. Typecheck, tests, build and deploy all passed, and the live site still shows the old text. Where do you look first?
The speed score changes every time you measure
One Lighthouse run said 99. Four more runs of the same build said 92, 94, 93 and 92. You need a number you can trust before deciding whether a change helped. What do you do?
Let's talk
Tell me what you're hiring for, or ask anything. It ends with an email you can read before you send it.
Your email app should have opened with this message. If it didn't, copy it and send it to forchadiwork@gmail.com.
Or directly
forchadiwork@gmail.com
linkedin.com/in/chadi-choukkah-802066177
github.com/shad-yy