𝐒𝐞𝐧𝐢𝐨𝐫 𝐀𝐈 𝐄𝐧𝐠𝐢𝐧𝐞𝐞𝐫 — 𝐋𝐋𝐌 𝐀𝐩𝐩𝐥𝐢𝐜𝐚𝐭𝐢𝐨𝐧𝐬 (𝐏𝐲𝐭𝐡𝐨𝐧, 𝐀𝐖𝐒 𝐁𝐞𝐝𝐫𝐨𝐜𝐤, 𝐑𝐀𝐆, 𝐕𝐮𝐞.𝐣𝐬) | 𝐑𝐞𝐦𝐨𝐭𝐞

We're hiring one engineer to own an AI assistant that's already in use and take it the rest of the way to production.
Teachers use our K-12 platform to plan lessons, assign work, and grade. Our assistant lives inside it: it answers from our curriculum and support content, levels articles to a student's reading grade, generates assessment questions, helps grade, and completes real tasks for teachers with their confirmation. The foundation is built; we need someone to own it and ship.
𝐖𝐡𝐚𝐭 𝐲𝐨𝐮'𝐥𝐥 𝐨𝐰𝐧
- The service end-to-end: Python, FastAPI, and LangGraph agents on AWS Lambda, with streaming
- Model routing on Bedrock (Nova, Claude, gpt-oss), balancing accuracy against cost per turn
- Retrieval: Pinecone hybrid search, OpenAI embeddings, reranking and index sync pipelines
- Evals for routing, retrieval, and answer quality, so every change is backed by numbers
- Safety for schools: Bedrock Guardrails, grounded answers, link allowlists, confirmation before data changes
- The in-app side in Nuxt 3 / Vue 3 / TypeScript, where the assistant reads the page and takes actions. Much of the feature work lands here
- Production: Terraform, CloudWatch alarms, cost tracking, GitLab CI
𝐖𝐡𝐚𝐭 𝐰𝐞'𝐫𝐞 𝐥𝐨𝐨𝐤𝐢𝐧𝐠 𝐟𝐨𝐫
- 5+ years building production software, strong Python
- 2+ years shipping LLM features to real users (RAG, agents, tool calling)
- Hands-on with a vector database and embeddings, including changing embedding models
- Solid AWS: Lambda, API Gateway, IAM, S3, DynamoDB; Bedrock a big plus
- Strong Vue.js and TypeScript: you'll build much of the frontend yourself
- MongoDB, Terraform, and CI/CD
- Fluent with AI coding tools, and reviews their output as carefully as a colleague's
𝐇𝐨𝐰 𝐰𝐞 𝐰𝐨𝐫𝐤
- Ownership: when something breaks, you find out why and fix it
- Small, reviewable increments every week, not one big drop a month
- Evidence: measure before you argue
- Say early when something is blocked or slipping
- Try a cheaper model or smaller prompt before a bigger one
𝐍𝐢𝐜𝐞 𝐭𝐨 𝐡𝐚𝐯𝐞
Laravel/PHP (our platform API), EdTech or FERPA/COPPA, LangSmith/Langfuse, LLM eval design, cost and latency optimisation
𝐖𝐡𝐲 𝐣𝐨𝐢𝐧
Real ownership of a product teachers use daily, a modern stack, competitive pay, fully remote.
📩 Apply: https://lnkd.in/dKRG4MYb

Work arrangement
Yes