Senior Software Engineer

Who you are

• 5+ years building production backend systems; deep comfort with Python, relational stores, and cloud infra (we're on GCP/Kubernetes)
• You've built LLM-powered systems in production — an agent loop, tool-calling integration, RAG, or eval infrastructure — and have opinions from the scar tissue
• You design for operability: legal compliance, tracing, evals, and kill switches are part of the feature, not afterthoughts
• You can own a technical domain with light supervision: propose direction in writing with clear API specs, take review, ship, and carry the pager for what you shipped
• Strong written communication — our operating model runs on design docs and PR review across team boundaries
• Nice to have: experience being the platform side of a platform/product relationship; SSE/streaming systems; compliance-adjacent engineering (data minimization, auditability); prior Staff scope

What the job involves

• Syndi is Syndio's AI platform — one shared agent experience (in-product chat, Slack, Teams) serving our product lines, built on syndi-api: an agent runtime owning orchestration, tool calling, memory, RAG, evals, and observability, live in production today
• The platform is young (v1 shipped this quarter), moving fast, and designed around a clear operating model: product teams contribute domain content through APIs, evals, and knowledge — the platform owns the experience
• We're hiring a staff-level engineer to take ownership of major runtime surfaces and grow into a technical owner of the platform
• This is a high-autonomy role on a small team (3–4 engineers): you'll design, ship, and operate systems end-to-end, and you'll work directly with product-team owners through the platform's contribution seams
• Core runtime surfaces of syndi-api: the agent loop, tool execution and registry, memory, context management (Python, Postgres, Claude on Vertex, GCP)
• The eval system — offline gates and online detector/judge evals written onto production traces — and its growth as product teams adopt it
• Production operation: observability (Datadog LLM Obs), incident response, the reliability of an agent surface real customers use
• The platform's contribution seams: reviewing product teams' tool wrappers and evals, evolving the GET /tools registry and drift-detection contract
• Clarifying, designing, and implementing compliance and legal requirements for Syndi
• First 30 days: Focus on onboarding, mastering our core agent runtime surfaces, and shipping a small-scale contribution to understand the platform's contribution seams
• First 60 days: Independently manage core runtime surfaces, troubleshoot complex issues, and start co-owning a consumer team's integration from API design to eval sign-off
• First 90 days: Concretely ship a major runtime capability end-to-end (e.g., multimodal input or a new channel surface), take on-call for the platform, and co-own a consumer team's integration from API design to eval sign-off

The application process

• Below you'll find an outline of the interview plan for our Senior Software Engineer (AI Platform) position. Please note that this is what we expect the process to look like; we may ask you for supplemental information or require an additional step before making a final decision
• Recruiter Screen (25 min)
• Hiring Manager Interview (30 min)
• Technical Evaluation - Design Skills (60 min)
• Crossfunctional Team Interviews (2 x 30 min)

Benefits

• Best medical plans on the market (90% employee coverage, 50% dependents)
• Up to 16 weeks of paid parental leave
• Life & Disability paid by the company
• 401(k) option
• Stock options
• Access to Talkspace online therapy