Senior Software Engineer — Platform & Infrastructure

Zohorecruit; Polska; Hybrid; PLN 260,000 - 380,000; Full time

Job summary

Conflation Labs is seeking a Senior Software Engineer for Platform & Infrastructure to own the RAG afig platform, multi-tenant architecture, and CI/CD pipelines. You will mentor a small team delivering AI coding agents and partnering with the CTO on architecture and strategic POCs.

You will design scalable architectures, drive platform generalization to new domains, and ensure robust monitoring and production readiness across tenants and city-scale deployments.

Qualifications

• 6+ years building production software with significant senior IC time.
• Hands-on RAG experience with production-grade retrieval, chunking, recall, reranking, eval drift.

Responsibilities

• Design with the CTO and turn POCs into shippable architectures.
• Own RAG+agentic platform and scale agents end-to-end.
• Maintain multi-tenant architecture with tenant isolation and data separation.
• Own CI/CD and testing infrastructure; ensure deployment quality.
• Develop AI Studio—self-service tenant onboarding and RAG config.
• Implement tenant-level monitoring, observability, and alerting.
• Triage and fix production issues; mentor juniors orchestrating coding agents.

Skills

RAG

Multi-tenant SaaS

DevOps

AWS

CI/CD

AI coding agents

Code review

Golang/Python/JS

Tools

Docker

GitHub Actions

Terraform

Ansible

PostGIS

Job description

Senior Software Engineer — Platform & Infrastructure

Senior IC role, reports to the CTO. Not a people-manager role. You partner with the CTO on architecture, scope new POCs and requirements into shippable work, and own delivery end-to-end. You own our RAG + agentic platform, our multi-tenant architecture, and the CI/CD pipelines that ship them. You mentor a small team of junior engineers who deliver primarily through AI coding agents — they orchestrate Claude Code and similar tools rather than writing most code by hand, and your job is to make that team's output good. THE CHALLENGE AHEAD Two parallel scaling paradigms will shape what you build:

• Breadth — many tenants. We're going from a handful of cities to 20+. Early customers ship in forward-deployed mode (bespoke per city); your job is to spot the patterns and pull common functionality back into a shared platform so city N+1 isn't a rebuild. This is also where "AI Studio" — our self-service portal for tenant onboarding, user management, and RAG configuration — fits: tenant setup as product surface, not engineering work per customer.
• Depth — large single-customer builds. Expect occasional high-volume engagements that won't fit the multi-tenant platform and will run as standalone systems with their own architecture. Learnings flow back into the platform over time. We also expect to generalize our RAG platform to domains beyond zoning — building code, fire, environmental, permitting. The platform abstraction that makes this possible is yours to own. Deciding what stays bespoke and what becomes platform — across all of these tracks — is part of the role.

WHAT THIS ROLE IS NOT

• Not a people manager — no 1:1s, no performance reviews, no headcount.
• Not pure architecture — the CTO owns the broader picture; you co-design and own execution.
• Not pure solo IC — you lead a small team's output, not just your own.

Requirements

Key Responsibilities

Design with the CTO. Take POCs and new requirements, turn them into shippable architectures, decompose them into tasks for the team.

Own our RAG + agentic platform. Build, scale, and debug our RAG agents (site-plan agent, zoning agent, and the next ones). Find failure modes, fix them, ship new agents end-to-end. Generalize the platform to support domains beyond zoning.

Own multi-tenant architecture — tenant isolation, configuration, data separation. Make city N+1 a configuration change, not a rebuild.

Own CI/CD and delivery. Design, maintain, and scale the pipelines that ship features to staging and prod. Build testing infrastructure (integration, load, regression). Make sure features land with the right tests, pre-deploy checks, and post-deploy validation.

Develop "AI Studio" — a self-service portal for tenant onboarding, user management, and RAG configuration.

Implement tenant-level monitoring, observability, and alerting.

Triage and debug production issues; deliver incremental fixes for tenant-specific problems.

Mentor juniors who orchestrate coding agents. Our juniors don't write much code by hand — they scope work, prompt agents, review output. You raise their bar through code review, design discussion, and patterns for getting good results from agents.

In this role, you must have

6+ years building production software, with significant senior IC time.

Hands-on RAG experience — you have built, scaled, and debugged Retrieval-Augmented Generation systems in production. You know what breaks (chunking, retrieval recall, reranking, eval drift, agentic loops) and how to fix it.

Experience building or maintaining multi-tenant SaaS systems — tenant isolation, configuration, data separation.

Strong DevOps fundamentals — Docker, CI/CD pipelines, deploy automation, AWS (EC2/RDS/S3), GitHub Actions.

Comfortable orchestrating AI coding agents as a daily practice — and comfortable mentoring others to do the same.

Shipped LLM-backed product features at scale: prompt caching, evaluation, observability, cost tracking, graceful fallbacks.

Strong production debugging skills and a production-first mindset.

You might thrive in this role if you have

Familiarity with geospatial indexing and queries (PostGIS). We build agentic systems that convert natural language to spatial queries.

Familiarity with observability tooling — Prometheus, Grafana, Sentry, CloudWatch.

Experience with GitOps and infrastructure-as-code (Terraform, Ansible).

Experience with vector databases at scale — pgvector, embeddings management, retrieval tuning.

Experience with RAG eval / observability tooling (Ragas, LangSmith, DeepEval).

About Conflation Labs

We build AI-powered software for Planning department, AEC firms and local government.

Our product turns stacks of PDFs — zoning ordinances, site plans, fire codes — into structured, queryable knowledge that helps planners and architects move faster.

Multi-tenant, live in production for cities including Asheville and Berkeley.

Small team, high ownership, close collaboration with the CTO and senior engineers.