Kairais • Careers
Build AI systems that hold up in production.
We're a small team building conversational AI and supporting systems with production standards. We value clear ownership, measurable outcomes, and calm execution. If you enjoy building reliable software under real constraints, you'll fit in.
Focus
Conversational AI • Retrieval • Applied Systems
Bar
Engineering standards and accountability
Open roles
24
How we work
- Clear scope, written acceptance criteria, and measurable outcomes
- Reliable systems: logging, testing, and incident-ready operations
- Secure-by-default practices and careful data handling
- Documentation that supports fast onboarding and clean handoffs
Artificial IntelligenceMachine LearningWeb DevelopmentApplication DevelopmentData Science
Open roles
Filter by level, location, type, and tech. Click a role to expand.
Build FastAPI services that are reliable in production: clean contracts, predictable performance, and strong observability.
Build maintainable Spring Boot services with strong security defaults, clean architecture, and production-grade operations.
Build reliable computer vision pipelines using OpenCV and modern model tooling, focused on performance, edge cases, and deployability.
Design conversation-first UX that feels trustworthy and controllable: clear system behavior, good recovery paths, and agent workflows users understand.
Build reliable pipelines for analytics and ML datasets using modern orchestration, transformations, and data quality checks.
Drive measurable improvements in conversational AI by designing experiments, analyzing user behavior, and building evaluation that holds up in production.
Build clean, fast product interfaces in React and Next.js with strong UX discipline and reliable API integration.
Develop backend systems in modern Java with attention to correctness, performance, and operational reliability.
Build and maintain API endpoints and backend features in Python while learning production habits: testing, logging, and clean contracts.
Analyze conversation data and product metrics to identify failure patterns, measure improvements, and support model and UX iterations.
Build clean UI components and pages in React/Next.js with good UX defaults and predictable state handling.
Support conversational AI development by running experiments, preparing datasets, and improving evaluation quality with strong guidance from senior ML engineers.
Own PostgreSQL reliability and performance for production systems, including schema design, query tuning, migrations, and operational health.
Develop PyTorch-based LLM and retrieval systems with strong evaluation discipline and reliable production deployment.
Build automation that prevents regressions across web UI, APIs, and chat flows, with stable CI execution and clear release gates.
Build secure, cross-platform desktop applications in Rust and Tauri with strong systems discipline and a focus on reliability.
Own application security across services, APIs, and supply chain: threat modeling, secure defaults, and automated security gates in CI.
Own LLM safety and adversarial resilience across RAG and tool-using agents, with measurable safety regressions and runtime guardrails.
Own retrieval and RAG quality end-to-end: data to indexing to reranking to evals, with production-grade latency and reliability.
Build the platform and operational discipline that lets ML and product teams ship safely: CI/CD, observability, environments, and incident readiness.
Own retrieval quality and latency for RAG: indexing, hybrid search, reranking, and measurable relevance improvements with production constraints.
Set the technical direction for conversational AI systems across retrieval, orchestration, evaluation, and safety, turning prototypes into reliable product capabilities.
Own platform architecture for a high-availability ConvAI product: deployment patterns, observability standards, cost control, and production governance.
Build and deploy TensorFlow models that power conversational AI features, with clear evaluation, stable training pipelines, and production-grade inference.