Integrant is looking for game changers to join our team as " Lead AI Platform".
About the role
The Lead AI Platform Engineer is responsible for bridging AI workloads with production-grade infrastructure, with a strong focus on NVIDIA AI stack, enabling high-performance, scalable, and optimized AI systems.
This role focuses on model optimization, runtime efficiency, and GPU utilization, ensuring that AI workloads are production-ready, cost-efficient, and performant across enterprise environments.
Roles and Responsibilities:
•
Translate AI/ML workloads into optimized infrastructure and deployment strategies
•
Optimize model performance across GPU environments (latency, throughput, memory utilization)
•
Design and implement inference and training pipelines using NVIDIA stack tools (TensorRT, Triton, NIM)
•
Convert and optimize models across frameworks (PyTorch → ONNX → TensorRT)
•
Analyze and resolve performance bottlenecks using profiling tools (GPU, memory, network)
•
Improve GPU utilization and scheduling efficiency across clusters
•
Design scalable distributed training and inference architectures
•
Work closely with customers to define AI infrastructure strategies and deployment models
•
Support production deployments including monitoring, rollback, and performance validation
•
Conduct applied research to improve model efficiency and infrastructure utilization
•
Mentor team members on AI infrastructure, optimization, and GPU systems
•
Experiment tracking tools (MLflow, W&B, Neptune) log parameters, metrics, and artifacts for comparison
•
Find the Model degradation happens post-deployment: concept drift, data pipeline changes, traffic pattern shifts
•
Root cause analysis (RCA) applies to ML systems: isolating variables, reproducing issues
Requirements
•
8+ years of experience in AI systems
•
8+ years of experience in ML systems, HPC and AI infrastructure
•
Strong proficiency in Python
•
Strong experience with GPU-based AI workloads and performance optimization
•
Deep understanding of model optimization techniques (quantization, pruning, batching)
•
Hands-on experience with:
1.
PyTorch
2.
ONNX / ONNX Runtime
3.
TensorRT / TensorRT-LLM
4.
Triton Inference Server
•
Knowledge of CUDA, cuDNN, and GPU architecture fundamentals
•
Experience with distributed systems (multi-GPU / multi-node)
•
Familiarity with:
1.
NCCL communication
2.
NVLink / InfiniBand
3.
Kubernetes or Slurm for orchestration
•
Experience deploying AI models into production environments
•
Ability to analyze system bottlenecks (compute, memory, network)
•
Experience with profiling tools (Nsight, TensorRT profiler, etc.)
•
Knowledge of cost optimization strategies for GPU workloads
•
Experiment tracking tools (MLflow, W&B, Neptune) log parameters, metrics, and artifacts for comparison
•
Find the Model degradation happens post-deployment: concept drift, data pipeline changes, traffic pattern shifts
•
Root cause analysis (RCA) applies to ML systems: isolating variables, reproducing issues
Nice to Have
•
Experience with NVIDIA NIM and NGC ecosystem
•
Exposure to Megatron-LM, NeMo, or large-scale LLM training/inference
•
Experience with LLM optimization techniques (KV cache, batching strategies)
•
Familiarity with MLOps practices and CI/CD for AI systems
•
Experience in customer-facing architecture or consulting roles
•
Familiarity with hybrid cloud / on-prem HPC environments