About PubNub
PubNub powers the world’s most engaging real-time experiences—chat, live updates, and interactive applications—for over 2,000 companies including Verizon, Autodesk, Zillow, and Dropbox. Our global data network processes trillions of messages each month with sub-100 ms latency across 15+ data centers. Backed by $130M in funding, we’re shaping the future of how the world connects.
About the Role
- Design and build AI-powered features for the Illuminate analytics platform
- Architect natural language to query systems that translate user intent into optimized SQL
- Build React-based interfaces for AI-driven insights, anomaly detection, and query builders
- Develop low-latency pipelines processing analytical workloads across ClickHouse and Postgres
- Create and maintain prompt architectures for complex multi-step AI workflows
- Build validation and error handling systems between LLM outputs and downstream data pipelines
- Integrate with Kafka/Kafka Streams for real-time data processing and AI inference
- Collaborate on system design with product and engineering teams across timezones
About You
Mandatory Requirements
- Production AI Experience: 1+ year shipping LLM-powered applications to customers (not just prototypes)
- Systems Engineering: Strong backend fundamentals with 5+ years experience in building high-throughput, distributed systems
- Database Expertise: Strong SQL skills, particularly analytical patterns (aggregations, window functions, CTEs, query optimization)
- Frontend Skills: Solid React experience building complex, data-intensive UIs
- LLM Engineering:
- Experience with structured output extraction (function calling, JSON mode, tool use)
- Ability to design and iterate on prompt architectures for complex tasks
- Understanding of validation patterns between AI outputs and downstream systems
- Systematic approach to testing non-deterministic systems
- Tech Stack: TypeScript, Kotlin, React
- Collaboration: Works effectively async across timezones; available 7am-12pm PT overlap
- AI-Assisted Development: Regular use of Copilot, Cursor, Claude, or similar tools
- Language: Fluent English
Nice to Have
- Experience with MCP (Model Context Protocol)
- Background with columnar/analytical databases (ClickHouse, BigQuery, Snowflake)
- Kafka or Kafka Streams experience
- Query planning and optimization at the database level
- Experience building evaluation datasets and regression test suites for LLM applications
- Knowledge of RAG patterns for schema/documentation retrieval
- Exposure to semantic parsing or pre-LLM NL-to-SQL approaches
- Understanding of caching and performance optimization for LLM-based systems
- Docker/Kubernetes experience
- Cloud AI platforms (AWS Bedrock, Azure OpenAI, Vertex AI)
- Prisma ORM experience
Why Join PubNub
- Build a greenfield intelligence platform at internet scale
- Ship features that land directly in customer-facing products across healthcare, fintech, gaming, and streamingAbout the Role
- High ownership and influence over technical decisions
- Remote-first setup within Poland
- Competitive B2B compensation ranging between PLN26,000 - PLN32,000 per month
- Optional office in central Katowice
If you're excited to build AI systems that make complex data accessible and actionable, we'd love to hear from you.
Similar Jobs
What We Do
PubNub powers apps that bring people together in real-time for remote work, play, learning, and health. Thousands of companies use PubNub's developer platform and APIs as the foundation for their online chat, virtual events, geolocation, remote control, and real-time updates, at a massive global scale. Since 2010, PubNub has invested in the tools and global infrastructure required to serve customers like Adobe, DocuSign, and RingCentral, delivering SOC 2 Type 2 security and reliability while meeting regulatory needs like HIPAA and GDPR. PubNub has raised over $130M from notable investors like Raine Group, Sapphire, Scale, Relay, Cisco, Bosch, Ericsson, and HPE.
