Senior Java Developer

What does Ozgar do?

Ozgar is an AI "Smart Knowledge Hub": it parses complex, poorly documented systems and turns code, databases and architecture (IBM, Cobol and Java) into structured, visual documentation and dependency graphs. Live since summer 2025, patent-backed, already sold to double-digit enterprise customers internationally. The hard part is the engineering: serving reliable, source-linked answers over decades of legacy code through a hybrid retrieval stack (vector + Apache AGE + Elasticsearch), without hallucination.

Team setup

Senior-heavy team — two Scrum teams, 90%+ seniors, with a product lead who has built systems at Amazon, Snowflake and IBM.

What you will do

• You develop and operate Ozgar across the full cycle: analysis, design, code, testing, production.
• You work on the hybrid-retrieval layer (vector embeddings, Apache AGE, Elasticsearch) and the LLM integration (Claude, OpenAI; on-prem models optional).
• You build adapters that ingest RPG/COBOL, Java/.NET into knowledge hub.
• You build agents for data ingestion and assistant.
• You build MCP / API connectivity to other agents.
• You evaluate new models and methods and bring the ones that hold up into the product.
• Over time you move from feature work into architecture ownership in your area.

Your profile

You have built and shipped backend systems in Java for several years and grown into senior or tech-lead scope. You have owned features end-to-end and already worked hands-on with LLM APIs in something real.

Must-have:

• Java 21 with Spring Boot, REST, Hibernate
• Solid CS fundamentals: OOP, data structures, concurrency
• PostgreSQL
• Hands-on with Claude or OpenAI APIs (integration, prompt engineering) and a working understanding of how LLMs behave — the AI layer is central, not a side feature

Nice to have:

• Neo4J, Elasticsearch, MongoDB, Kubernetes
• LangChain4J / Spring AI, open-source models (Qwen, gpt-oss)
• Language: English at upper-intermediate or better — the team works in English day to day.

Hiring process

1. Recruiter call

2. Technical interview with the Tech Lead, ~60min

3. Meet the Product Lead, ~60 min

4. Offer

End-to-end realistically ~2–3 weeks.