Hello
I'm Alex Sandruk.
Senior / Lead Applied AI Engineer with backend and platform depth.
I build AI integration layers, backend services, internal tools, and delivery loops that teams can understand, verify, and operate after the first demo is over.
My best work sits where applied AI meets real software pressure: product workflows, platform boundaries, observability, browser QA, GitOps, handoffs, and the human decision gates that keep fast work from becoming fragile work.
Make the work observable before you make it clever.
What I do
I build AI integration layers, internal tools, and backend/platform systems where the hard part is not only making something work once, but making it inspectable, recoverable, and owned by a real team.
The strongest fit is Senior or Lead Applied AI Engineer work with backend and platform depth: APIs, orchestration, workflow tools, runtime checks, observability, CI/CD, browser verification, and human approval loops around AI output.
- Turn agentic or AI-assisted prototypes into reviewable product workflows.
- Connect tools, data, prompts, browser automation, and delivery checks without losing ownership boundaries.
- Make fast-moving technical work easier to explain, test, ship, and return to later.
Tech I work with
My strongest current stack is TypeScript/Node.js, Python/FastAPI, backend services, PostgreSQL, Playwright/browser automation, Cloudflare, CI/CD, Docker, and LLM/API integration.
Earlier work in Solidity, Go, DevOps/SRE, product operations, and fintech/blockchain systems still informs how I think about auditability and delivery discipline.
- TypeScript, Node.js, NestJS, Python, FastAPI, PostgreSQL.
- LLM APIs, agents, workflow tooling, structured outputs, eval/check loops.
- Playwright, browser QA, GitOps, Cloudflare, CI/CD, Docker, runtime verification.
How I work
I use AI aggressively, but I do not treat model output as authority. The useful loop is still human: source of truth, owner, next check, evidence, and a decision someone can stand behind.
I like small slices that can be inspected: a runnable change, a test result, a browser check, deployment evidence, a rollback path, or a note that lets the next session continue without replaying the whole conversation.
- Make vague work observable before optimizing it.
- Prefer boring reliability over impressive demos nobody can maintain.
- Separate what is proven, what is a strong guess, and what still needs a check.
Roles that fit
The cleanest public label is Senior or Lead Applied AI Engineer with backend and platform depth. Depending on the company, the same work can be called AI Integration Engineer, Senior Backend/Platform Engineer for AI-heavy products, Developer Experience Engineer with AI focus, or Technical Lead for AI workflow systems.
I am not positioning for pure ML research, model training, content-only evangelism, or demo-only prompt engineering roles without engineering responsibility.
- Best: product/platform teams turning AI capabilities into production workflows.
- Good: fintech, blockchain, developer tools, infrastructure, automation, or operations-heavy domains.
- Possible: hands-on technical leadership or engineering-manager-adjacent work when implementation depth still matters.
Beyond the code
Writing, podcasts, communities, and research are part of the engineering loop for me. They keep me close to the language people use when a tool is confusing, overhyped, promising, or finally useful.
A system is only partly technical. The other part is whether people can understand it, trust it, and return to it when the first context is gone.
- I like explaining complex technical work without flattening the hard parts.
- I care about re-entry: docs, handoffs, context, and recovery paths.
- I stay close to how developers actually adopt tools, not only how tools are marketed.
Few useful pages