SOFTWARE · CLOUD · AI — FULL-STACK ENGINEER
I build productsthat ship.
McDarsene M Mwale — full-stack engineer and founder of Mvalex, a Shanghai-based technology studio. I design, build, and operate digital products end-to-end: from the first commit to production traffic.
~/mcdarsene %
About
Engineer first. Studio second.
I'm McDarsene M Mwale, a full-stack engineer and the founder of Mvalex, a Shanghai-based technology studio. I work where software, cloud, and AI meet, building products end to end, from the first schema to the on-call rotation.
My work spans web applications, internal dashboards, mobile apps built with React Native and Flutter, and cloud platforms with CI/CD and observability. I design AI systems retrieval pipelines, agents, tool-use workflows — and I write the prompts, evals, and guardrails that make them reliable in production.
I care about the boring parts: typed interfaces, reproducible builds, observability, and cost discipline. Most of the leverage in software comes from doing the unglamorous things well, on schedule, for a long time.
Through Mvalex I take on a small number of engagements a year founding engineering, platform builds, and AI integrations for teams that need senior hands, not headcount.
Selected Work
Things I've built
Engagements
How I can help
Five practice areas covering the full lifecycle of a digital product — from the first commit to the production deploy and beyond.
Software Engineering
End-to-end product work — architecture, implementation, testing, and delivery for web, dashboard, and mobile.
- Full-stack web apps (Next.js, React, Node)
- Internal dashboards and admin tools
- Mobile apps (React Native, Flutter)
- API design and integration
Cloud Engineering
Cloud platforms designed for cost, reliability, and the team that has to run them.
- Cloud architecture and infrastructure as code
- Infrastructure as code (Terraform, CDK)
- CI/CD pipelines and release automation
- Observability, alerting, and cost tuning
AI & ML Engineering
AI systems that survive contact with production — retrieval, agents, evals, and guardrails.
- Retrieval-augmented generation (RAG)
- LLM agents with tool use and human-in-the-loop
- Fine-tuning, distillation, and eval harnesses
- Vector stores and semantic search
Automation
Removing operational toil — the workflows that should never have needed a human.
- Business process automation
- Integration between SaaS tools and internal systems
- Scheduled jobs, webhooks, and event pipelines
- Data sync and reconciliation
Prompt Engineering & AI UX
Prompts as interfaces — versioned, tested, and reviewed like any other production artifact.
- System prompt design and evaluation
- Chain-of-thought, tool use, and structured output
- Prompt regression testing and A/B evaluation
- Human-in-the-loop workflows and guardrails
0+
Years shipping code
0+
Repos on GitHub
0+
CI/CD workflows
0+
Mobile apps
Capabilities
The stack I build on
Four disciplines I keep sharp — covering the surface, the systems, the motion, and the persistence layer of a modern product.
Frontend
Typed, tested interfaces — web apps, dashboards, and marketing sites that feel fast because they are.
TypeScript · React · Next.js · Tailwind · Framer Motion
Backend & Data
APIs, schemas, and pipelines designed to be changed safely by people who aren't you.
Node · Postgres · Drizzle · tRPC · dbt · Airflow
Cloud & Platform
Infrastructure that fails predictably, recovers automatically, and costs what the invoice says.
Terraform · Docker · GitHub Actions · Vercel · EAS
AI & Automation
LLM systems and workflow automation — from prompts and evals to agents that run unattended.
Claude · OpenAI-compatible · GitHub Actions · TypeScript
FAQ
Questions, answered
What kinds of engagements do you take?
Founding engineering, platform builds, and AI integrations. Most projects run 6–16 weeks. I work solo or with a small team from Mvalex, depending on scope.
Do you work with existing engineering teams?
Yes — often. I embed as a senior IC or tech lead, ship alongside the team, and leave behind runbooks, ADRs, and tests.
How do you handle AI projects?
Same as any other software: clear problem, measurable outcome, versioned prompts, an eval harness, and a fallback path for when the model gets it wrong.
What's your stack?
TypeScript and React on the front-end, Node and Postgres on the back-end, GitHub Actions for CI/CD, and Claude or OpenAI-compatible models for LLM work. React Native or Flutter for mobile, depending on the team.
Do you build mobile apps?
Yes — React Native and Flutter. I pick based on the team that will maintain it, not personal preference.
Where are you based and what time zones do you work with?
Shanghai. I keep overlap with Asia-Pacific daily and with Europe and North America by arrangement.
How do projects start?
A 30-minute call to understand the problem, then a short written proposal — scope, milestones, and price. No discovery theatre.
What does a project cost?
Engagements typically start around $12k for a focused build. Fixed-scope, milestone-billed. I'll tell you if the budget and the goal don't match.
Contact
Let's build something that ships.
I take on a small number of engagements each quarter — founding engineering, platform builds, and AI integrations. Tell me what you're working on and I'll tell you honestly whether I'm the right fit.
What's your name?
So I know who I'm talking to.
No payment now. We confirm goals and next steps on the first call.