AI/ML platform Engineer

Role & responsibilities

1. LLM/SLM Engineering

• Design, fine-tune, optimize and deploy Large Language Models / Small Language Models using

  LoRA, QLoRA, and full-scale training

  approaches.
• Work hands-on with

  HuggingFace

  ,

  PyTorch

  , and internal training frameworks for custom model development.
• Train, benchmark and evaluate models such as

  Llama, Mistral, Qwen, Gemma, GPT-J, Falcon

  , etc.
• Build scalable pipelines for model evaluation, tokenization, dataset preparation and finetuning.
  

2. Data Engineering

• Develop and maintain robust

  Python-based data pipelines

  for preprocessing and cleaning large-scale datasets.
• Work with

  Pandas / Polars

  for data manipulation, transformation and analysis.
• Ensure data integrity, efficiency, lineage and reproducibility across model training cycles.
  

3. Model Deployment & Optimization

• Optimize inference using

  ONNX Runtime, TensorRT, GGML

  , and quantization techniques (

  int8/int4

  ).
• Deploy ML models to production environments with high availability, performance tuning, and cost optimization in mind.
  
• Build containerized inference systems using

  Docker

  and orchestrate deployments on

  Kubernetes

  clusters.
• Implement caching, batching, and hardware acceleration strategies for faster inference.
  

4. DevOps / LLMOps

• Own end-to-end model lifecycle management including versioning, registries, traceability, and rollback.
  
• Configure CI/CD pipelines for automated model validation, testing, deployment and monitoring.
  
• Monitor inference latency, throughput, GPU utilization, and infrastructure cost footprints.
  
• Build internal systems for model serving, A/B testing, performance dashboards, and real-time observability.
  

5. Collaboration & Technical Execution

• Work closely with Data Engineers, Platform Engineers, Infra, and Research teams to build stable, scalable AI systems.
  
• Conduct POCs, architecture reviews, and capacity planning for large-scale LLM deployments.
  
• Ensure engineering best practices, code quality, security compliance, documentation and reproducibility.
  
• Stay updated with latest advancements in LLM optimization, distributed training, vector DBs, and compute efficiency.