What you'll work on:
Integrate and productionize LLM-based systems: from prompt registries and model gateways to agentic workflows – with a focus on reliability, observability, and cost efficiency.
Own our core MLOps infrastructure: build and maintain our feature store, vector database, training and monitoring tools, etc.- so DS and product teams can ship AI features without infrastructure becoming the bottleneck.
Build AI-powered Python microservices that integrate LLMs, scraping, pipelines and internal data into product features used by hundreds of thousands of businesses.
Collaborate deeply with data scientists, engineers and product managers to turn ideas into reliable, scalable ML systems, iterating quickly from prototype to production.
Own initiatives end-to-end: from design doc to production deployment, including monitoring, documentation, and handoff.
Here is what is needed:
4+ years of production engineering experience, with at least 3 years working on data-intensive systems in Python.
Mindset of ownership, curiosity, and can-do attitude in a fast‑moving environment.
Hands-on experience with LLM-based systems in production – prompt management, scale, cost optimization, latency, and reliability.
Distributed systems fluency – event-driven architecture, idempotency, backpressure, schema evolution.
Workflow orchestration experience, both batch and streaming.
Comfortable with ambiguity – scoping unclear requests, making speed vs. robustness tradeoffs, communicating transparently when things change.
Strong collaboration and communication skills in Hebrew and English
Nice to have:
Experience with MLOps tooling (feature stores, training pipelines, model serving, or equivalent).
SQL and data warehouse fluency (DBT, Snowflake, or equivalent).
