AI Infrastructure/Software Engineer(GPU Computation)

We Have 2 different requirement below is the JD's for both the requirement:

ROLE-1 : AI Infrastructure Engineer (GPU Computation)

ROLE 2:AI Software Engineer (Generative AI)

1. AI Infrastructure Engineer (GPU Computation)

Experience Required: 5+ years in AI/ML infrastructure or high-performance computing

Location: Delhi NCR

Infrastructure Environment

Hybrid both cloud (e.g. AWS) and on-premises GPU clusters

Key Responsibilities

• Design, configure, and optimize GPU compute pipelines on-premises and cloud-based to hit target utilization levels (~90%).
• Profile training and inference workloads to identify and remove hardware-level bottlenecks, regardless of where they're hosted.
• Recommend and implement scaling, batching, and parallelization strategies across on-prem and cloud GPU clusters.
• Decide, workload by workload, whether on-premises or cloud compute is the better fit, and manage the two consistently.
• Collaborate continuously with the AI Software Engineer to optimize model performance, compute utilization, deployment architecture, and experimentation speed.
• Set up monitoring and reporting on GPU utilization, training throughput, and compute efficiency metrics across both environments.
• Identify opportunities to improve platform infrastructure capacity planning, cost efficiency, reliability ahead of being asked.
• Support a hypothesis-driven experimentation culture by enabling fast, low-friction infrastructure for rapid prototyping and iteration.

Success Metrics

• GPU utilization against target (~90%).
• Training/inference throughput and experiment turnaround time.
• Infrastructure reliability (uptime, incident frequency).
• Cost optimization across cloud and on-premises spend.

Requirements

• 5+ years of hands-on experience in AI/ML infrastructure or high-performance computing.
• Strong expertise in GPU architecture and tooling (e.g., NVIDIA CUDA, TensorRT).
• Experience with distributed training/inference frameworks (e.g., DeepSpeed, Horovod, Ray).

ROLE-2

2. AI Software Engineer (Generative AI)

Experience Required: 5+ years building and deploying generative AI / LLM-based solutions

Location: Delhi NCR

Key Responsibilities

• Design, build, and fine-tune generative AI models and pipelines for business use cases, covering the full lifecycle — prompt engineering, RAG implementation, agentic workflows, model evaluation, guardrails, and production monitoring.
• Collaborate continuously with the AI Infrastructure Engineer to optimize model performance, compute utilization, deployment architecture, and experimentation speed.
• Foster a hypothesis-driven experimentation culture by rapidly prototyping, measuring outcomes, and iterating based on feedback.
• Evaluate and benchmark internal AI solutions (e.g., EAP) against commercial offerings such as Microsoft Copilot and Excel Insights, to identify gaps and recommend improvements.
• Identify opportunities to improve AI platform capabilities, usability, and adoption through continuous experimentation and innovation.
• Incorporate feedback from retrospectives to continuously enhance model and platform performance — and close the loop visibly with the client.
• Apply LLMOps/MLOps practices for model deployment, versioning, and monitoring in production.

Success Metrics

• Experiment turnaround time.
• AI feature/tool adoption rate across business functions.
• Model accuracy and reliability against agreed benchmarks.
• User satisfaction with delivered AI capabilities.
• Number of successful production deployments.

Requirements

• 5+ years of hands-on experience building and deploying generative AI / LLM-based solutions.
• Strong experience with generative AI frameworks and tooling (e.g., LangChain, LangGraph, LlamaIndex, Hugging Face, OpenAI/Anthropic APIs).
• Experience with RAG pipelines, vector databases (e.g., Pinecone, Milvus, FAISS), and Model Context Protocol (MCP) or similar agentic/tool-use integration patterns.
• Familiarity with AI evaluation frameworks and building in guardrails for production AI systems.
• Working knowledge of FastAPI, Docker/Kubernetes, and CI/CD practices for AI workloads.
• Proven experience productionizing AI/ML models for real-world business use cases.
• Ability to work in a fast-paced, experiment-driven environment with tight feedback loops.