AI/ML Engineer

Key Responsibilities:

LLM Deployment & Optimization

• Deploy, fine-tune, and optimize open-source LLMs (e.g., LLaMA, Mistral, CodeS, DeepSeek).
• Implement quantization (e.g., 4-bit, 8-bit) and pruning for efficient inference on commodity hardware.
• Build and manage inference APIs (REST/gRPC) for production use.

Infrastructure Management

• Set up and manage on-premise GPU servers and VM-based deployments.
• Build scalable cloud-based LLM infrastructure using AWS (SageMaker, EC2), Azure ML, or GCP Vertex AI.
• Ensure cost efficiency by choosing appropriate hardware and job scheduling strategies.

MLOps & Reliability Engineering

• Develop CI/CD pipelines for model training, testing, evaluation, and deployment.
• Integrate version control for models, data, and hyperparameters.
• Set up logging, tracing, and monitoring tools (e.g., MLflow, Prometheus, Grafana) for model performance and failure detection.

Security, Compliance & Performance

• Ensure data privacy (FERPA/GDPR) and enforce security best practices across deployments.
• Apply secure coding standards and implement RBAC, encryption, and network hardening for cloud/on-prem.

Cross-functional Integration

• Work closely with AI solution engineers, backend developers, and product owners to integrate LLM services into the platform.
• Support performance benchmarking and A/B testing of AI features across modules.

Documentation & Internal Enablement

• Document LLM pipelines, configuration steps, and infrastructure setup in internal playbooks.
• Create guides and reusable templates for future deployments and models.

Key Requirements:

Education:

• Bachelors or Masters in Computer Science, AI/ML, Data Engineering, or related field.

Technical Skills:

• Strong Python experience with ML libraries (e.g., PyTorch, Hugging Face Transformers).
• Familiar with LangChain, LlamaIndex, or other RAG frameworks.
• Experience with Docker, Kubernetes, and API gateways (e.g., Kong, NGINX).
• Working knowledge of vector databases (FAISS, Pinecone, Qdrant).
• Familiarity with GPU deployment tools (CUDA, Triton Inference Server, HuggingFace Accelerate).

Experience:

• 4+ years in an AI/MLOps role, including experience in LLM fine-tuning and deployment.
• Hands-on work with model inference in production environments (both cloud and on-prem).
• Exposure to SaaS and modular product environments is a plus.