AI Inference Engineer / NVIDIA GPU AI Engineer

Company: Tesseract Energy

Location: Chandigarh, India

Employment Type: Full-time

Travel: Periodic travel to Australia required

Reporting To: Tesseract Australia Executive and Platform Leadership Team

About Tesseract Energy

Tesseract Energy is building Australia's first distributed AI inference platform, delivering secure, reserved, and managed GPU compute capacity to enterprise, telecommunications, AI software, and managed-service customers.

Our platform aggregates GPU compute facilities — from battery-backed edge pods equipped with NVIDIA RTX PRO 6000 Blackwell servers to larger commercial and regional data centres — and presents them as a unified, Australian-hosted inference service.

The platform combines:

- Reserved and on-demand GPU capacity (firm, flexible, burst, and batch)
- Managed private inference endpoints with strict tenant isolation
- Distributed Kubernetes and NVIDIA GPU orchestration (MIG, DCGM, GPU Operator)
- Regional capacity pooling, failover, and disaster recovery
- Battery-supported power resilience (BESS) and energy-aware workload scheduling
- Australian data residency and enterprise-grade security

Tesseract has an established office in Chandigarh supporting sales and marketing. We are now expanding this office to build the software and product capability required to deliver the platform in partnership with our Australian team.

Job Summary

We are looking for a skilled AI Inference Engineer with hands-on experience in NVIDIA GPU computing, AI model deployment, and inference optimization. The candidate will be responsible for deploying, optimizing, and scaling AI/ML models on GPU-based infrastructure to achieve high throughput, low latency, and effective GPU utilization.

The ideal candidate should have a strong understanding of NVIDIA GPU architecture, CUDA programming fundamentals, AI inference frameworks, model optimization techniques, and production-grade AI serving systems.

Key Responsibilities

Design, develop, and optimize AI inference pipelines using NVIDIA GPU platforms.

Deploy and serve machine learning and deep learning models on GPU infrastructure.

Optimize inference performance using NVIDIA TensorRT, CUDA, and other GPU acceleration technologies.

Implement model serving solutions using NVIDIA Triton Inference Server, TensorRT-LLM,



or similar frameworks.

Optimize GPU utilization, memory consumption, latency, and throughput for inference workloads.

Work with Large Language Models (LLMs), computer vision models, and other deep learning workloads.

Implement model quantization techniques such as FP16, BF16, INT8, and INT4, where supported and appropriate.

Profile GPU workloads and troubleshoot performance bottlenecks using NVIDIA Nsight Systems, Nsight Compute, and related tools.

Develop and maintain scalable inference services using REST APIs, gRPC, Docker, and Kubernetes.

Collaborate with AI/ML engineers, infrastructure teams, and software developers to integrate optimized inference solutions into production.

Monitor GPU performance, inference latency, request throughput, and system resource utilization.

Evaluate emerging NVIDIA GPU technologies and inference optimization techniques.

Required Technical Skills

1. NVIDIA GPU & CUDA

Strong understanding of NVIDIA GPU architecture and GPU computing fundamentals.

Knowledge of CUDA programming, GPU kernels, memory hierarchy, and parallel execution.

Experience with GPU memory management, profiling, and performance optimization.

Familiarity with NVIDIA data center GPUs, such as A100, H100, H200, or newer architectures.

2. AI Inference & Model Optimization

Hands-on experience deploying deep learning models for inference.

Strong knowledge of inference optimization techniques, including:

Model quantization and precision optimization.

Dynamic and static batching.

Memory optimization and efficient GPU utilization.

Kernel fusion and operator optimization.

Latency and throughput tuning.

Understanding of the trade-offs between model accuracy, inference speed, and hardware utilization.

3. NVIDIA AI Software Stack

Experience with one or more of the following technologies:

NVIDIA TensorRT

NVIDIA Triton Inference Server

TensorRT-LLM

CUDA and cuDNN

NVIDIA Nsight Systems and Nsight Compute





NVIDIA NIM and related inference deployment tools

4. AI/ML Frameworks

PyTorch and/or TensorFlow.

ONNX and ONNX Runtime.

Hugging Face Transformers.

Experience deploying LLMs, vision models, or other deep learning models.

5. Programming & Infrastructure

Robust programming skills in Python and/or C++.

Experience building inference APIs using FastAPI, Flask, REST, or gRPC.

Working knowledge of Linux, shell scripting, and system debugging.

Experience with Docker and containerized deployments.

Familiarity with Kubernetes, GPU scheduling, and cloud or on-premises GPU infrastructure.

Preferred Qualifications

Experience optimizing LLM inference, including KV-cache management, continuous batching, and multi-GPU inference.

Knowledge of distributed GPU computing and multi-GPU communication using NVIDIA NCCL.

Familiarity with vLLM, SGLang, or similar high-performance inference engines.

Experience with GPU performance benchmarking and capacity planning.

Understanding of GPU virtualization, MIG, and GPU resource allocation.

Experience deploying AI inference workloads in production environments.

Familiarity with monitoring tools such as Prometheus and Grafana.

Key Performance Indicators (KPIs)

The candidate will be evaluated on the ability to:

Reduce model inference latency.

Increase inference throughput and concurrent request capacity.

Improve GPU utilization and memory efficiency.

Optimize cost per inference or cost per token.

Maintain inference service availability and reliability.

Deliver repeatable performance benchmarks and production-ready deployments.

Educational Qualifications

Bachelor's or Master's degree in Computer Science, Computer Engineering, Electrical Engineering, AI/ML, or a related field.

Relevant industry experience and demonstrable GPU optimization expertise may be considered in place of advanced academic qualifications.

Ideal Candidate Profile

The ideal candidate is an engineer who understands both AI models and the underlying GPU execution environment. They should be able to take a trained model, deploy it on NVIDIA GPU infrastructure, profile its performance, identify bottlenecks, and optimize it for production workloads.

Pay: ₹800,000.00 - ₹1,400,000.00 per year

Work Location: In person

📌 AI Inference Engineer / NVIDIA GPU AI Engineer (Manauli)
🏢 Tesseract EnergyAustralia
📍 Manauli