AI Engineer (RAG & FineâTuning Specialist)

Location:

Type:

Experience:

Qualification:

Responsibilities

• Architect & implement the RAG pipeline: embeddings ingestion, vector search (MongoDB Atlas or similar), and context-aware chat generation.
• Design and build Python‑based services (FastAPI) for generating and updating embeddings.
• Host and apply LoRA/QLoRA adapters for per‑user fine‑tuning.
• Automate data pipelines to ingest daily user logs, chunk text, and upsert embeddings into the vector store.
• Develop Node.js/Express APIs that orchestrate embedding, retrieval, and LLM inference for real‑time chat.
• Manage vector index lifecycle and similarity metrics (cosine/dot‑product).
• Deploy and optimize on AWS (Lambda, EC2, SageMaker), containerization (Docker), and monitoring for latency, costs, and error rates.
• Collaborate with frontend engineers to define API contracts and demo endpoints.
• Document architecture diagrams, API specifications, and runbooks for future team onboarding.
  

Required Skills

• Strong Python expertise (FastAPI, async programming).
• Proficiency with Node.js and Express for API development.
• Experience with vector databases (MongoDB Atlas Vector Search, Pinecone, Weaviate) and similarity search.
• Familiarity with OpenAI’s APIs (embeddings, chat completions).
• Hands‑on with parameters‑efficient fine‑tuning (LoRA, QLoRA, PEFT/Hugging Face).
• Knowledge of LLM hosting best practices on AWS (EC2, Lambda, SageMaker).
  

Containerization Skills (Docker)

• Good understanding of RAG architectures, prompt design, and memory management.
• Strong Git workflow and collaborative development practices (GitHub, CI/CD).
  

Nice‑to‑Have

• Experience with Llama family models or other open‑source LLMs.
• Familiarity with MongoDB Atlas free tier and cluster management.
• Background in data engineering for streaming or batch processing.
• Knowledge of monitoring & observability tools (Prometheus, Grafana, CloudWatch).
• Frontend skills in React to prototype demo UIs.