Overview
Sembix Strata™ is our flagship enterprise semantic-modeling platform: workspace agents, delivery planning, and telemetry deployed into regulated public sector environments where the customer's security team reads every line of our architecture diagram. This is not a services role. You will not rotate across client engagements. You will work directly with the CTO on one codebase in a two-person-deep loop, and you will still be maintaining the decisions you make eighteen months from now.
We are looking for one person: a senior engineer who is unusually careful. Someone who reads the diff twice, who notices the case that only breaks under concurrency, and who gets a small thrill out of chasing a bug down to the actual mechanism instead of the first plausible story. Strata runs agents that write to shared state on behalf of customers. A silent normalization that quietly downgrades a plan, a torn read between two handlers writing the same file, a stale zone record that blocks certificate issuance - these are not theoretical to us. They are the last few months. The work is interesting precisely because the failure modes are subtle and the blast radius is real.
About Sembix
Sembix is redefining how software gets built, through AI-native product engineering. We take systems from idea to production—designing, building and rewriting applications into cloud-native systems. Our solutions help organizations unlock the full potential of the cloud, driving efficiency and innovation.
Responsibilities
- Design, build, and own end-to-end features in Strata: workspace agents, delivery planning, telemetry, and the platform surfaces around them.
- Work through architecture with the CTO as a genuine counterpart - expected to disagree, propose alternatives, and sometimes be the reason we don't ship something.
- Own correctness in the concurrent, distributed parts of the product: shared filesystem state, multi-handler writes, agent orchestration, idempotency, and recovery.
- Debug hard problems to root cause - across the application, the container runtime, and AWS infrastructure - and write down what you found so the next person doesn't rediscover it.
- Build the evaluation and observability tooling that tells us whether an agent behavior actually improved, rather than felt better in a demo.
- Raise the reliability bar: tests that would have caught the last incident, guardrails that make a class of bug impossible, and release hygiene that makes deploys boring.
- Contribute to accessibility (WCAG 2.2 AA), security review responses, and the compliance posture the product is held to.
- Occasionally sit with customer engineering teams during deployment and diagnose issues live.
Required Qualifications
- 5+ years building and operating production software, with meaningful time owning a product rather than delivering projects.
- Deep TypeScript; strong Node and React/Next.js; comfortable in Python where it's the right tool.
- Real experience with concurrency and distributed state: races, idempotency, transactional boundaries, eventual consistency, and the debugging techniques that surface them.
- Hands-on AWS: containers on ECS/Fargate, IAM, networking, and Bedrock; can read a CloudWatch trace and Terraform plan without hand-holding.
- Practiced with CI/CD, containerization, and infrastructure-as-code as part of your own workflow, not someone else's.
- Shipped LLM-powered features to production users; formed opinions on prompt and context design, retrieval, tool use, agent orchestration, and how to measure it.
Preferred Skills
- Public sector, GovTech, EdTech, or another regulated domain.
- Developer tooling, IDE- or workspace-style products, or anything involving code analysis at scale.
- ASTs, static analysis, and legacy codebase comprehension.
- SOC 2, NIST 800-53, FedRAMP, or similar compliance exposure.
- pgvector, Postgres at depth, OpenTelemetry, or an AWS certification.
Success Profile
- You check the boundary case and finish the last 10% - error states, empty states, docs - before calling something done.
- You want to know why a bug happened, not just that it stopped.
- You write design notes and PR descriptions a colleague can follow six months later.
- You'll tell the CTO when you think he's wrong, and commit properly once a call is made.
- You can tell us about a bug that took you more than a day: what you initially thought was wrong, what was actually wrong, and how you found the difference.
- You've shipped something with an LLM in it that behaved badly in production, changed it, and knew it worked.
Compensation & Benefits
- Base salary: $135,000 - $160,000.
- Incentive compensation: 17% of base salary, tied to product delivery and quality outcomes.
- Total potential compensation: approximately $158,000 - $187,000.
- 401(k): 100% match up to 4% of base salary, fully vested immediately.
- Paid Time Off: 20 days per year.
- Medical: 100% employee coverage and 50% family coverage.
- 10 paid holidays annually.
Why Join Sembix
- Be part of a mission-driven team shaping the future of education and public service.
- Work on high-impact products that simplify delivery and improve public-sector outcomes.
- Flexible, remote-friendly work culture with a focus on results.
- Competitive compensation and benefits.
- Opportunity to grow into leadership roles as we scale our impact.
How to Apply
Send your resume and answers to the two questions below. Short is fine; specific is essential.
- Tell us about a bug that took you more than a day to solve.
What did you initially think was wrong, what was actually wrong, and how did you find the difference?
- Tell us about something you shipped with an LLM in it that behaved badly in production.
What did you change, and how did you know it worked?
We read every answer. The second question tells us a lot about your experience with production AI. The first tells us how you think when the obvious explanation turns out to be wrong.
Apply through the Sembix careers portal: https://sembix.ai/careers.html
Equal Opportunity
Sembix is an Equal Opportunity Employer. We are committed to creating a diverse, inclusive, and respectful workplace where all employees and applicants are treated fairly.
Employment decisions at Sembix are based on business needs, job requirements, qualifications, and merit. We do not discriminate on the basis of race, color, religion, sex, gender identity or expression, sexual orientation, pregnancy, national origin, age, disability, genetic information, veteran status, marital status, or any other status protected by applicable federal, state, or local law.
Sembix is committed to providing reasonable accommodations for qualified individuals with disabilities and for sincerely held religious beliefs, practices, or observances, unless doing so would create an undue hardship.