LLM Infrastructure & Deployment Summer Unpaid Internship

To promote and accelerate whole brain emulation, Carbon copies Foundation focuses on encouraging scientific collaboration, publishing reviews and maintaining research roadmaps. We aim to increase general awareness of the human adaptability gap and the potential of neural prostheses, whole brain emulation, and substrate‑independent mind.

We are seeking an LLM Infrastructure Intern with a Dev Ops mindset to build, deploy, and maintain the production systems that house our large‑scale neural models. While researchers focus on the "what," you will focus on the "how"—ensuring our LLM pipelines are scalable, observable, and cost‑effective. This is a remote, unpaid summer internship ideal for a software engineer wishing to master the deployment lifecycle of foundation models within a mission‑driven neuroscience organization.

• Model Deployment & Orchestration:
  Deploy LLMs and embedding models using containerization (Docker/Kubernetes) and orchestration tools such as Bento
  
  ML, Ray Serve, or vLLM.
• CI/CD for LLMs:
  Build and automate continuous integration/continuous deployment pipelines for model updates, ensuring seamless integration between research code and production environments.
• Vector Database Management:
  Administer and optimize production vector databases (Pinecone, Milvus, or Qdrant) for high‑performance retrieval in RAG systems.
• Monitoring & Observability:
  Implement logging and monitoring for LLM performance, tracking metrics like Tokens Per Second (TPS), latency, and model drift using tools such as Weights & Biases or Lang Smith.
• Infrastructure as Code (IaC):
  Use Terraform or Ansible to manage cloud‑based GPU clusters and storage solutions for neural data.
• API Development:
  Develop and maintain high‑performance FastAPI/Flask wrappers around model endpoints for use by the wider Carbon copies developer community.

• Python Mastery:
  Strong experience in backend Python development with asynchronous programming and API design.
• Dev Ops Toolkit Proficiency:
  Docker and version control (Git). Experience with Kubernetes or cloud providers (AWS, GCP, Azure) is a major plus.
• LLM Serving:
  Familiarity with inference optimization techniques such as quantization (GGUF/EXL2), caching strategies, and load balancing.
• Database Knowledge:
  Experience managing relational (Postgre
  
  SQL) and No
  
  SQL/Vector databases.
• Systems Thinking:
  Ability to troubleshoot complex distributed systems and optimize resource allocation for GPU‑intensive tasks.
• Educational Background:
  Currently pursuing a degree in Computer Science, Software Engineering, or a related technical field.

• Flexible, approximately 20 hours per week; internship duration 3–4 months.

The Carbon copies Foundation is committed to diversity and welcomes applications from all qualified individuals regardless of race, color, religion, gender, sexual orientation, national origin, age, disability, or veteran status.