Job Description & Details
This is a solid hands-on engineering role focused on building out modern generative AI workflows and deploying LLM applications into production. If you are tired of just writing toy prompts and want to tackle real-world orchestration, RAG architectures, and model maintenance, this gig actually gets into the weeds of production AI.
What You'll Actually Be Doing
You will spend your days bridging the gap between raw data science concepts and stable, production-ready AI pipelines. Expect to write heavy Python scripts for data preprocessing, fine-tune prompt workflows, and wire up APIs that feed into LLMs. You will also deal with the unglamorous side of AI—containerizing models with Docker, setting up monitoring in production, and collaborating closely with data scientists to make sure their experimental models don't crash when real users hit them.
The Core Tech Stack
You need to be fluent in Python and comfortable navigating modern generative AI tooling like LangChain, Hugging Face, and LlamaIndex. Having a solid grasp of Docker is non-negotiable since you're helping push these models to production. Familiarity with RAG systems, hybrid search techniques, and multi-agent workflows will instantly put you at the top of the pile, as the team relies heavily on these patterns to keep their AI accurate and contextual.
Interview Expectations
Expect the hiring team to ask you to debug a failing RAG pipeline where retrieval latency is spiking or hallucinations are creeping in; they want to see how you troubleshoot vector embeddings and context window limits under pressure. You will also likely face a system design question around containerizing a PyTorch or TensorFlow model with Docker and scaling it efficiently behind a REST API, so make sure you understand the performance bottlenecks of serving large language models.
Application Advice
Skip the generic objective statement and highlight your actual production experience with Python, Docker, and LLM frameworks right at the top of your resume. Make sure to explicitly mention terms like prompt engineering, RAG, and workflow orchestration so you clear the ATS keyword filters, and briefly describe a project where you took an AI model from a local Jupyter notebook and actually shipped it somewhere useful.