AI Engineer – AI‑First Backend & LLM Specialist

About the role

We are seeking an AI Engineer who embraces an AI‑first mindset and has a solid backend development background. You will design, build, and operate production‑grade AI systems where intelligence is baked into the core architecture rather than added as an afterthought.

Key responsibilities

• Develop scalable backend services with Python frameworks such as FastAPI, Flask, or Django to support AI‑driven applications.
• Implement, fine‑tune, and deploy large language models (e.g., GPT‑4, Claude, Llama 3, Mistral) using Hugging Face, PyTorch, or TensorFlow.
• Design advanced Retrieval‑Augmented Generation pipelines and integrate vector databases like Pinecone, Weaviate, Milvus, PGVector, or Chroma.
• Build AI agents and multi‑agent workflows with LangChain, LlamaIndex, Google ADK, BAML, Agno, or CrewAI.
• Own the full MLOps lifecycle: experiment tracking, CI/CD deployment, monitoring for hallucinations, and latency/cost optimization.
• Architect AI‑native systems that prioritize semantic search, embeddings, and context‑aware data flows.
• Deploy and optimize open‑source LLMs on‑premise using serving engines such as vLLM, SGLang, or Ollama.

Required profile

• Minimum 4 years of professional Python experience, including async programming, type hinting, and high‑performance backend patterns.
• Deep understanding of transformer architectures, attention mechanisms, and tokenization.
• Hands‑on experience with vector embeddings and similarity search optimization.
• Proficiency in building AI data pipelines, data cleaning, ingestion, and prompt engineering.
• Strong DevOps/MLOps background with Docker, Kubernetes, and cloud platforms (AWS, GCP, Azure). Experience with GPU orchestration and model quantization (GGUF, AWQ) is a plus.

Required skills

• Python
• FastAPI, Flask, Django
• Transformers, LLMs (GPT‑4, Claude, Llama 3, Mistral)
• Hugging Face, PyTorch, TensorFlow
• RAG pipelines
• Pinecone, Weaviate, Milvus, PGVector, Chroma
• LangChain, LlamaIndex, Google ADK, BAML, Agno, CrewAI
• vLLM, SGLang, Ollama
• Docker, Kubernetes
• AWS, GCP, Azure
• GPU orchestration, model quantization