AI Engineer (Backend-Focused, Node.js + Python)

We’re looking for an AI Engineer to join our outsourcing team — not a pure ML researcher, but a pragmatic, data-driven backend engineer who knows how to put LLMs, vision APIs, and off-the-shelf models to work inside real products.

Most of our work isn’t training models from scratch — it’s understanding the business logic, collecting and cleaning data, wiring it into LLMs or services like AWS Textract, and shipping the result as a reliable API inside our existing platform.

If you’ve spent the last few years building APIs and integrations, you’re comfortable with data, and you’ve leaned hard into Cursor or Claude Code to deliver projects faster — this is for you.

What you’ll do

• Build and ship LLM-powered features: prompt design, data prep, passing context to the model, parsing and validating output, returning clean APIs.
• Design and implement integrations between our product and third-party AI services (OpenAI, Anthropic, Hugging Face models, AWS Textract, etc.).
• Take document-processing and computer-vision tasks end-to-end: ingest → Textract/vision API → post-process → structured output consumed by our backend.
• Write backend services and APIs in Node.js (our primary backend), dropping into Python when a task calls for it (model work, data scripts, fine-tuning).
• Fine-tune existing models and prompts — not from scratch, but iterating on what’s already in place to get better accuracy, cost, or latency.
• Do the unglamorous but critical work: data collection, cleaning, transformation, pattern-finding — often by letting an LLM do the heavy lifting on structured and unstructured input.
• Maintain and improve existing AI and automation workflows already running in production.
• Monitor cost and performance of model/API usage, propose optimizations, and forecast costs for new features.
• Prepare and present demos and solution overviews to U.S.-based clients.
• Document architecture, prompts, and pipelines in Confluence.

What we’re looking for

• 2+ years shipping production backend systems, with meaningful hands-on experience integrating LLMs or AI APIs into real products.
• Strong in Node.js with the flexibility to work in Python — or a Python engineer who’s comfortable committing Node.js code. We need someone fluent on both sides; pure-Python candidates won’t be a fit.
• Solid grasp of algorithms and data structures — enough to write custom logic when an LLM or library can’t cover the case.
• Real experience with LLM integrations: OpenAI, Anthropic, or similar. Prompt engineering, function/tool calling, structured output, retrieval, evaluation.
• Familiarity with the Hugging Face ecosystem — picking models, running inference, fine-tuning where it makes sense.
• API and integrations background: REST, auth, webhooks, rate limits, retries, idempotency — the stuff that makes integrations not break at 2 a.m.
• Comfortable with data work: collecting, cleaning, transforming, and analyzing data (SQL, pandas, or the Node equivalents).
• Working knowledge of AWS — especially Textract, Lambda, S3, and the services around document and data pipelines. Bonus for Athena, Glue, Kinesis, SageMaker.
• Exposure to NLP and Computer Vision concepts — enough to pick the right tool and evaluate results, not necessarily to train SOTA models.
• Familiarity with CI/CD and deploying AI-powered services to production.
• Experience building chatbot workflows — conversational flows, multi-turn logic, context management — is a plus, as this is among our planned projects.
• AI-native delivery workflow. You use Cursor or Claude Code daily and know how to ship a project with minimal hand-written code — scoping work for the agent, reviewing generated code critically, orchestrating multi-step tasks, and getting to shipped quality faster than a traditional workflow. If you can’t describe how you use these tools to deliver, this role isn’t the right fit.

Nice to have

• Experience fine-tuning open-source LLMs or vision models on domain-specific data.
• Background in fintech, healthcare, or construction domains.
• Experience with vector databases, RAG pipelines, or agentic workflows.

What we offer

• Competitive salary with regular reviews.
• Support for individual entrepreneurs with accounting assistance.
• Fully remote work.
• A clear path for professional development and growth.
• A friendly team environment and engaging team-building activities (online and offline).