Senior Technical Architect

$60–130/hrrate as listed

Review and score AI-generated system designs, architecture decision records, and multi-service code changes for correctness, scalability, and maintainability. Work is remote and asynchronous through micro1, with pay observed in the $60–130/hr range depending on stack depth and assessment results.

Summary, eligibility guidance and screening preparation are prepared by Sidequest from the platform's posting. Check the source listing before applying.

What the work involves

You will act as the senior reviewer standing between a model's output and a plausible production system. Typical tasks include grading two competing architecture proposals for the same requirements document, writing reference solutions for distributed-systems prompts (idempotency, backpressure, cache invalidation, partition strategy), critiquing AI-authored ADRs where the reasoning sounds fluent but the tradeoff is wrong, and annotating multi-file diffs that touch service boundaries. Some projects ask for adversarial prompt authoring: design a scenario where a model will confidently recommend the wrong consistency model, then document why.

Most assignments are delivered as batches with rubrics. You score against dimensions like requirement coverage, failure-mode handling, operational realism, and cost awareness — and your written justification matters as much as the number. Reviewers who only mark "good"/"bad" without naming the specific failure mode do not get renewed on projects.

What the platform screens for

An AI-led interview that probes one system you actually designed, then follows up on the parts you glossed over — expected traffic, what broke first, what you would change now.
Evidence of depth in at least one primary stack plus real exposure to cloud infrastructure, data stores, and CI/CD, rather than breadth-only familiarity.
Written clarity under time pressure; micro1 weighs the quality of your explanations heavily because the artifacts become training signal.
Consistency between your resume claims and your live answers — vague seniority claims are a common rejection reason.

Logistics

Fully remote, asynchronous, contractor engagement.
Commitments commonly land between 10 and 20 hours per week; some projects offer more during ramp-ups.
Deadlines are per-batch rather than per-hour, so overlap with a specific time zone is rarely required, though occasional sync calibration calls happen.
Pay is reported in the $60–130/hr band, set by assessed level and project; rates are as observed on the platform and not guaranteed.

What they look for

Cloud Benchmark Task Authoring
Cloud and Distributed Systems Architecture
Production Infrastructure Ownership

Before you apply

Would you pass this screen?

Drawn from what this listing actually requires. Answered in your browser — nothing is sent to us, and we do not store your answers.

Have you personally designed and shipped an architecture for a production system that served traffic for at least six months?
Production system architecture · required
Do you have hands-on experience configuring, deploying, or troubleshooting systems on at least one public cloud platform (AWS, GCP, Azure) or equivalent managed infrastructure?
Cloud infrastructure hands-on · required
Have you architected or significantly modified a data store strategy (choice of SQL, NoSQL, caching layer, or partitioning schema) for a system in production?
Data store depth · required
Have you worked on a system that spans multiple services or nodes where you had to reason about failure modes, consistency, or cross-service communication?
Distributed systems experience · preferred
Can you provide a writing sample—design doc, ADR, postmortem, or technical blog post you authored—that demonstrates your ability to explain complex architecture decisions clearly?
Written technical communication · preferred

0 of 5 answered

Included with this listing

Interview prep — 11 questions

Matched to what this platform's screening actually measures. Answer each out loud in 60–90 seconds: claim → concrete example → implication.

What it tests: Whether you can articulate the actual business or technical pressure that shaped your decisions, not just the tech stack you chose. Screening for real architectural ownership versus inherited codebases.
A strong answer: Name a specific system, state the core constraint (scale, latency, cost, regulatory, or operational) that drove your design, and show how that constraint directly shaped a key architectural choice—not just list technologies used.
Where people lose it: Describing a system you worked on without claiming clear ownership of the architecture decision itself, or naming constraints that sound generic (scalability, reliability) rather than specific (5-minute P99 SLA at <$2/user/month, or data residency in EU only).