Note: The job is a remote job and is open to candidates in USA. TheFutureSociety is seeking an experienced Backend Engineer to support MAX, its AI astrology platform. The role focuses on LLM inference infrastructure, caching, retrieval, Postgres, token usage, entitlements, audit reporting, and server-side application logic through independently testable fixed-fee gates.
Responsibilities
- Run git log on the prompt-assembly file and tell us in writing when the current cache-breakpoint placement was introduced and by whom
- Token telemetry — reliable input/output and cache-token tracking
- Inference caching — restructure prompt caching to reduce unnecessary token costs. Conversation history currently sits inside the cached section of the prompt, causing the cache to be invalidated on new messages and the history to be rewritten at cache-write rates. We have measured the behavior and will provide the full specification
- Entitlements — server-side temporary tier grants and automatic trial expiry
- Inference audit & reporting — audit unnecessary generation calls and build per-tier usage reporting
Skills
- Experienced Backend Engineer with strong LLM/inference infrastructure experience
- Significant production backend experience
- Comfortable with LLM APIs, RAG, inference, caching, tokens, and prompt assembly
- Familiar with Postgres/Supabase and SQL
- Comfortable working from detailed technical specifications and independently tested acceptance criteria
- Contribute immediately and work confidently within a production LLM stack
Benefits
- Remote work
- Paid trial with potential for ongoing work
- Fixed fee per gate
- Async only, no recurring meetings
- Staging + read-only production access from day one
- PR-based workflow with required regression testing
- Existing diagnostics provided so you can start implementing immediately
Company Overview