Scope the real problem
Start from the outcome and the cost of failure. Most requests shrink once the actual problem is on the table; build the smallest system that solves it.
Software Engineer · Backend & AI Infrastructure
I build software that runs itself.
From multi-tenant APIs and event-driven backends to AI gateways and smart contracts: I take problems from scoping to shipped, with the automation that keeps them running unattended.
15+ client projects, scoped to shippedFreelance & contract · remote worldwide5.0 ★ · 100% job success on Upwork
Across the stack
Backend APIsEvent-driven systemsAI infrastructureSecurity & accessAutomation
What I build
Different systems. Same discipline.
APIs, AI gateways, smart contracts, bots — the surface changes, the job doesn't: make the rules explicit, put them in code, and let the software enforce them.
Bring me a problem →- Systems
Backend platforms & APIs
Multi-tenant services, REST and gRPC APIs, and event-driven workflows that stay clean as the product grows around them.
Go · TypeScript/Node · Postgres - Intelligence
AI behind guardrails
LLM gateways, grounded assistants and RAG pipelines — behind the policy, rate-limit and audit layers that make a model safe in front of paying users.
Up to 80% off the time to put a bespoke, interactive client proposal in front of a buyer.
Gateways · grounding · RAG · MCP - Trust
Security & access
Auth, RBAC, request signing, key rotation and contract audits — permission boundaries enforced at runtime, not in a wiki.
JWT/OAuth2 · RBAC · PoP signing · audits - Automation
Machines doing machine work
Event-driven bots, CI/CD, containerised deploys and alerting that take a person out of work a person shouldn’t be doing.
~24 hours of manual event processing out of a client’s week, after a repo-monitoring system moved onto Cloudflare Workers and SMTP pipelines.
CI/CD · Docker · Cloudflare Workers
Client reviews
The client's side of it.
Every review here comes from a completed Upwork contract — condensed for the page, with the originals checkable on the profile.
Uprising talent on Upwork5.0 average100% Job Success10+ projects delivered remotely
An exceptional full-stack engineer. He delivered beyond expectations — he didn’t just write code, he helped refine the logic and flow of the product itself. Professional communication and true ownership of the project. Highly recommend.
Work and beyond — definitely work with him again.
Great freelancer. I hired him to work on an open-source project and he delivered everything, exactly as scoped.
How I work
From vague ask to running system.
The work is not done when the code merges. It is done when the system is deployed, documented, and running without anyone watching it.
Put the rules in code
Permissions, pricing, rate limits, approvals: if a rule matters, it gets enforced at runtime — not remembered by a person or written in a doc nobody reads.
Ship it finished
Deployment, documentation, alerts, and handover go out with the code. Finished means it runs, and keeps running, with nobody in the loop.
Writing
Works, with the reasoning included.
Write-ups from shipping backends, AI infrastructure, and automation — what got built, what broke, and the tradeoffs behind the decisions.
A spec for coding agents needs precedence rules and a graveyard
Structure a large-migration spec so AI agents (or contractors) can execute it safely: explicit conflict-precedence rules, a rejected-alternatives section, phase gates with acceptance criteria, and a stop-and-flag escape hatch.
Deletion across data systems is a workflow, not a query
Coordinate erasure through an idempotent, checkpointed workflow when one subject's data spans operational, analytical, and AI stores.
Enforce budgets at the level the human approved them
Aggregate usage counters up to the granularity the approver actually saw, so wildcard expansion can't silently multiply an approved ceiling by the number of rows it fanned out into.
How to measure the real cost of an agent evaluation
A practical accounting model for judge calls, agent calls, retries, and the unattributed work that per-case dashboards often hide.
About
End to end, then out of the loop.
I'm a backend-focused engineer working in Go and TypeScript, with a software engineering degree and a minor in AI. I've shipped multi-tenant SaaS platforms, event-driven systems, smart contracts, and AI integrations — as a freelancer, at an agency, and as open source.
Clients rarely hand me a spec. They hand me a problem, and I take it end to end: scoping, architecture, build, deployment, documentation. Along the way, the rules that matter — permissions, pricing, rate limits, approvals — move out of people's heads and into the code.
The common thread is not a stack. It is that the system keeps working — and keeps enforcing its rules — after I have stepped away.
Get in touch
Is a person still doing work your software should?
Tell me what you're building, where the manual work or the fragile part lives, and what finished looks like. I'll tell you whether I can build it — and the shortest route if I can. Twenty minutes on a call is usually enough to know.
Prefer a platform contract? Hire me on Upwork ↗Connect on LinkedIn ↗