Software Quality Assurance Engineer

THE COMPANYMeridian is an AI operating partner for service businesses. We embed within our clients’ operations and deploy AI that runs core business workflows end-to-end: admissions, lead qualification, candidate screening, and client engagement. We connect their systems, we work with their data, and we build AI experiences tailored to how their business actually operates.This is a fundamentally different model from selling software or building tools. Our clients do not buy a product and figure out how to use it. We sit inside their operations, identify where AI creates the most leverage, deploy agents that handle real conversations with real people in English, Arabic, and Portuguese, and iterate on performance every week based on live data. We are measured on their outcomes: conversion rates, response times, cost per acquisition, placement speed.Our engineers design systems and direct AI to produce production software. We ship to production every week. We operate across the UAE, KSA, and Brazil with live enterprise clients in education, healthcare staffing, and professional services, and solutions that are already delivering measurable ROI.HOW WE WORKWhat follows are the operating conditions of this team. Read them carefully. If they do not describe the environment where you do your best work, this is not the right place for you.Customer problems first, technology second. We succeed because we sit inside our clients’ operations and translate their real problems into AI experiences that produce measurable results. Every product decision starts with a customer problem that we can name, quantify, and solve. If you cannot trace a feature back to a specific client outcome, it does not get built.​Ship every week or fall behind. We move experiments and new features to production every week. We maintain an intelligence layer across our systems that accelerates our development lifecycle and keeps the entire team aligned on what is live, what is next, and what is blocked. If a week passes without a production release, something is wrong, and we fix it immediately.​Meticulous organization is a weapon. We are obsessive about how we organize work. Every ticket has clear requirements, edge cases, and acceptance criteria. Every integration has documented dependencies. Every client deployment has a defined SOP. This discipline is what allows a small team to operate at the output level of a team five times its size.Candor over comfort. We tell clients what they need to hear. We do the same with each other. If something is not working, we say it directly, with respect, and we fix it. We do not protect people from honest feedback, and we do not expect to be protected from it. Consensus is not the goal. The right answer is the goal.Small team, no passengers. Every person on this team carries real weight. There is no layer of middle management to hide behind. You own your work end-to-end, ask for help when you need it, and deliver. If you want structure handed to you, this is not the right environment. If you want to build the structure, keep reading.Day 1, always. We have paying clients and real traction. None of that entitles us to anything. The market does not care what we built last quarter. Our competitors are funded, fast, and coming. The only advantage we have is that we are closer to our clients’ operations, we move faster, and we refuse to treat anything as settled. It is still Day 1. It will always be Day 1.WHAT YOU WILL OWN01 · Deployment pipeline ownership. You own the full path from dev to staging to production. You ensure that every change that is ready in dev gets promoted reliably and on schedule. You build and maintain the CI/CD infrastructure that makes deployments repeatable, auditable, and fast. When a fix is merged, it reaches production. That accountability is yours.02 · Production incident response. You are the first line of defense when something goes wrong in production. You triage, diagnose, and resolve issues or escalate to the relevant engineer with clear context. You own the monitoring and alerting that lets you catch problems before clients report them.03 · Release gating. You define and enforce the checklist that must be completed before any deployment reaches production. You validate that acceptance criteria are met, integrations are stable, and regressions have not been introduced. You have veto authority on releases that do not meet the bar and you work with the Product Manager to balance shipping speed with quality thresholds.04 · Automated test pipelines. You build and maintain CI/CD integrated test suites covering API contracts, webhook flows, prompt regression testing, and UI functionality. You instrument monitoring and alerting to catch production anomalies before clients report them.05 · AI-specific validation. You develop evaluation frameworks for LLM outputs: response quality scoring, hallucination detection, tone consistency, multilingual accuracy (English, Arabic, Portuguese), and adherence to client-specific SOPs and persona guidelines. Testing AI is fundamentally different from testing deterministic software, and you build the strategies that account for probabilistic outputs.06 · Infrastructure and deployment health. You manage the AWS infrastructure that our platform runs on. You own alerting, logging, uptime monitoring, and capacity planning. You ensure the infrastructure supports our weekly release cadence and scales reliably as we add clients.07 · Integration testing. Our agents connect to CRMs (HubSpot, Netsuite), telephony providers, WhatsApp Business API, and various client systems. You own the test coverage for these integration points, ensuring that upstream changes do not break downstream flows.WHAT WE ARE LOOKING FOR01 · Infrastructure and DevOps experience. Solid working knowledge of AWS (EC2, Lambda, S3, CloudWatch, or equivalent). You have managed production infrastructure, set up alerting and monitoring, and maintained deployment pipelines. You understand what it takes to keep a multi-tenant platform running reliably.02 · CI/CD and release engineering. You have built or maintained deployment pipelines in GitHub Actions, GitLab CI, or similar. You understand how to structure a dev to staging to production promotion workflow that is fast, reliable, and auditable. You have opinions about what belongs in a release checklist, and you enforce them.03 · QA engineering ability. At least 3 to 5 years in a QA, SDET, or release engineering role. Hands-on experience building automated test suites with tools like Playwright, Cypress, or similar. You understand the difference between test coverage that looks good on a dashboard and test coverage that actually catches bugs.04 · AI/ML product testing. Experience validating the outputs of language models, voice AI systems, or conversational AI platforms. You understand that testing an LLM requires evaluation strategies built for probabilistic outputs: tone scoring, hallucination detection, multilingual accuracy, persona adherence.05 · Programming ability. You write code in Python and/or TypeScript. You can read production code well enough to identify where bugs originate and write targeted tests against specific modules. You are comfortable with REST APIs, webhook debugging, and reading server logs.06 · Incident response temperament. You stay calm under pressure, you communicate clearly during an outage, and you write post-mortems that actually prevent recurrence. You have carried on-call responsibility before, and you know how to triage effectively.07 · Languages. Fluent in English. Portuguese and/or Arabic is a significant advantage.Strong plus: Experience with voice AI testing (STT/TTS pipelines, latency measurement, pronunciation accuracy), WhatsApp Business API testing, Arabic language NLP evaluation, or Terraform/infrastructure as code workflows.HOW TO APPLYSend a short note and your CV/portfolio/public repo to jose@meridiaventures.xyz. Tell us one thing you have tested or broken that taught you something valuable and why Meridian is interesting to you. We value directness over polish.Our process: an initial conversation with the CEO or Chief of Staff, a practical exercise relevant to the role, and a final conversation with the team. Start to offer in under two weeks for the right candidate.​