Skip to content
View Mrkewl's full-sized avatar

Block or report Mrkewl

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
Mrkewl/README.md

Jazsley Zainal

**Full-stack engineer | AI/RAG systems | GovTech-scale delivery |

I build production software that ships to real users — from national-scale government platforms serving millions of Singaporeans, to AI-powered apps I design, build, and use myself. My background spans exercise science, 6 years running a fitness business, and 4+ years in software engineering. That cross-domain perspective shapes how I approach technical problems: from the user backward, not the stack outward.


What I'm building

Out of the box, Claude is capable but undisciplined — it skips requirements, guesses intent, and loses context between sessions. Agent
Harness fixes that by enforcing structured thinking, planning, and execution as non-negotiable gates before any code gets written.

Forget about prompts. Just describe what you want — Agent Harness routes your request to the right workflow automatically. Starting a
feature, debugging a bug, shipping a branch — each one triggers the correct skill without you having to think about it.

Every session starts with project context loaded. Every feature goes through brainstorm → spec → plan → implement → review. Every agent follows best practices baked into its definition — not suggested, enforced.

It ships with a library of reusable agents (architect, backend engineer, code reviewer, team lead), a workflow skill for each phase of
development, a persistent knowledge base that grows with your project, and a

FitForge — AI fitness coaching grounded in sports science

RAG-based coaching app that surfaces peer-reviewed training research (Helms, Israetel, Schoenfeld) as personalised, evidence-backed recommendations. Not another generic AI wrapper — the domain knowledge is the product.

Flutter Koa/TypeScript Supabase pgvector Anthropic Claude API PostgreSQL

  • Built a full RAG pipeline: chunked sports science literature → pgvector embeddings → Claude API retrieval and generation
  • Grounded in Eric Helms' adherence pyramid, Mike Israetel's RP volume landmarks, and Brad Schoenfeld's hypertrophy research

ApplyAI — Job application pipeline, automated

End-to-end job search CRM I built to solve my own problem. Scrapes LinkedIn, Greenhouse, and Lever; scores JDs against my profile; generates tailored resumes and cover letters; auto-applies via Playwright. Currently in personal production use.

Next.js TypeScript Supabase Anthropic Claude API Playwright PDFKit Tailwind CSS

  • Claude API integration: JD scoring, resume tailoring, cover letter generation — all from a single profile.json source of truth
  • Custom PDFKit renderer: hanging-indent bullets, inline bold parsing, zero third-party PDF libraries
  • Kanban CRM with Supabase persistence, per-platform location normalisation, and staleness detection

Where I've shipped

Palo IT → GovTech Singapore (2023 – present) Built and maintained digital services used by millions of Singaporeans: LifeSG (cross-agency CMS), Baby Bonus System (30,000 babies/year), OneService App v2. Reduced development redundancy 50% via an internal Form Builder Library.

Next.js React Native Koa PostgreSQL AWS ConfigCat GitLab CI/CD

Accenture (2022 – 2023) SRE and cloud infra for enterprise systems. 95%+ uptime, AWS Lambda automation, Dynatrace/CloudWatch monitoring.

An Xing Technologies (2021 – 2022) Cross-platform Flutter apps, Stripe/Apple Pay/Google Pay integrations, mentored 4 NUS interns to full-time conversion.

Bodyengineers Pte Ltd — Founder (2015 – 2021) Built and ran a fitness business leasing personal trainers to national gym chains across Singapore. 200+ clients, press features in The New Paper and Berita Harian, invited speaker at Pepsi Company and en world Singapore. Received a 4/5 grant evaluation from an Enterprise Singapore accredited incubator for an AI personal training platform — before I knew how to code.


Stack

Layer Technologies
Frontend / Mobile React, Next.js, Flutter/Dart, TypeScript, Tailwind CSS
Backend Node.js, Koa, Django, Python
Databases PostgreSQL, Supabase, pgvector, Firestore, MongoDB
Cloud / DevOps AWS, Azure, Docker, Kubernetes, Terraform, GitHub Actions, GitLab CI/CD
AI / ML Anthropic Claude API, RAG pipelines, pgvector, prompt engineering
Observability CloudWatch

Certifications

  • AWS Solutions Architect – Associate (2022)
  • AWS Developer – Associate (2024)
  • Terraform Associate (2023)
  • Microsoft Certified: Azure Administrator (2023)
  • GitHub Copilot Certified (2025)
  • ICT Assessment (Software Developer) – GovTech Digital Academy / NUS-ISS (2025)
  • Anthropic: Claude 101, Claude Code in Action, Introduction to MCP (2026)

Education

B.Sc. Sports Science — Edinburgh Napier University Diploma in Aerospace Engineering — Singapore Polytechnic Certified Practitioner — Mac-Nutrition University


📍 Singapore | 🌍 Open to relocation: UK, Germany, Netherlands, Australia 📧 jazsleyzainal.93@gmail.com 💼 linkedin.com/in/jazz

Popular repositories Loading

  1. Cat-Dog-Classifyer Cat-Dog-Classifyer Public archive

    Dart

  2. Mrkewl Mrkewl Public

  3. My-Portfolio My-Portfolio Public

  4. My-Certifications My-Certifications Public

  5. Jaygents Jaygents Public

    TypeScript