Our client is not just another start up on the block. They are here with a grand mission to use AI to help humanity come together to solve complex problems. Their main goal is improving productivity by helping to streamline the decision making process.
Key Responsibilities:
What You’ll Be Rocking:
- Large Language Model System Engineering
- Design and build scalable LLM-based architectures that play nice with external data sources and APIs, with a spotlight on RAG systems.
- Monitor and optimize system performance, ensuring our clients get the best in class.
LLM Development:
- Craft intricate LLM workflows using multiple models, data sources, and processing steps.
- Implement CI/CD pipelines for swift changes, seamless deployment, and robust monitoring.
- Set up and maintain cloud-based infrastructure for LLM applications.
- Keep an eagle eye on LLM performance and resource utilization.
- Automate testing and QA processes to keep our LLMs sharp and reliable.
Debugging and Quality Assurance:
- Tackle issues head-on, improving LLM performance, cost efficiency, accuracy, and reliability.
- Use data analysis and experimentation to continuously fine-tune our systems.
- Cross-Functional Collaboration:
- Work hand-in-hand with product managers, engineers, and other stakeholders to gather requirements and deliver impactful solutions.
Your Superpowers:
- Mastery of LLM frameworks, especially with a focus on RAG (e.g., LLamaIndex, Langchain).
- Experience with Large Language Models (e.g., GPT-3, GPT-4, PaLM) and their APIs.
- Proven prowess in system design and building/deploying LLM-based applications.
- DevOps and LLMOps expertise, including CI/CD pipelines, cloud platforms, and monitoring tools.
Our Tech Playground:
- Frontend: VueJS, tailwind-ui, firebase auth, hosted on Vercel.
- Backend: Node, Express, Prisma, Postgres, hosted on Heroku.
Skills We’re After:
- AI frameworks: LLMs, NLP (e.g., spaCy), LangChain, Embeddings.
- Technologies: Python, Flask, JWT, Postgres, Prisma, Docker, Postman.
- Process: Agile Scrum, Git, Pull-Request reviews.
Many more details to share!