Description
Summary Job description We're looking for an AI/ML engineer to help us design and ship [your product or feature — e.g., an LLM-powered support agent]. You'll work directly with our team to take ideas from prototype to production and keep them reliable once they're live. What you'll work on Building LLM-powered features using APIs like OpenAI, Anthropic, or open models Designing RAG pipelines: chunking, embeddings, retrieval, and vector search Writing prompts and evaluation harnesses, then iterating based on real outputs Integrating models into our backend and deploying to [AWS / Azure / GCP] Adding monitoring, logging, and guardrails so quality holds in production What we're looking for Strong Python and solid engineering habits: version control, testing, readable code At least one LLM or ML feature you've shipped to production (not just notebooks) Familiarity with a framework like LangChain, LlamaIndex, or equivalent Experience with a vector database (pgvector, Pinecone, Weaviate, or similar) Comfortable with cloud deployment and API design Clear written English and the ability to explain technical trade-offs Nice to have Fine-tuning or model evaluation experience Frontend skills for building AI-facing UIs Domain experience in [e.g., healthcare, fintech, e-commerce] Project details Engagement: 100-150/hr Commitment: ~5 hrs/week, ongoing Duration: 2–3 months to start, with potential to exten Hours: some overlap with US Eastern for standups; otherwise flexible To apply Please include: A short note on a relevant AI/LLM project you've shipped and what your role was Links to code, demos, or write-ups (GitHub, live apps, case studies) Your answers to the screening questions below We read every proposal and reply through Upwork. Looking forward to seeing what you've built. Screening questions (add these in the "Questions" step — up to 3 works well): Describe an LLM or ML system you built end to end. What was the hardest technical problem, and how did you solve it? Which tools do you reach for when building a RAG pipeline, and why? What's your weekly availability and your timezone?