Platform Engineer Intern - GPU Infrastructure & AI Inference

At

Host360

, we are building an AI infrastructure platform that transforms GPU servers into a scalable, multi-tenant AI cloud. Our goal is to enable customers to share GPU resources efficiently, deploy AI models, serve inference requests and manage AI workloads through a unified platform.

We are looking for exceptionally curious engineering students and fresh graduates who enjoy building things, experimenting with open-source technologies and solving challenging systems problems.

This is an opportunity to work on real infrastructure at the intersection of

GPU virtualization, Kubernetes, distributed computing and large language model (LLM) inference

.

What you will work on

GPU sharing and slicing:

Explore GPU partitioning, GPU time-slicing, resource isolation and efficient allocation of GPU resources across multiple workloads.

Kubernetes-based AI infrastructure:

Work with Kubernetes, GPU device plugins, NVIDIA GPU Operator and related cloud-native technologies.

AI inference systems:

Deploy and benchmark open-source inference engines such as vLLM, NVIDIA Triton Inference Server or SGLang.

Distributed AI workloads:

Experiment with Ray and KubeRay for distributed execution, workload orchestration and scalable model serving.

Performance engineering:

Measure inference latency, throughput, GPU utilization, VRAM consumption and cost per inference request.

Platform engineering:

Help build automation, APIs, monitoring and control-plane components that make AI infrastructure easy to operate.

Research and experimentation:

Evaluate open-source projects, understand their internals, run experiments and turn promising ideas into working prototypes.

What we are looking for

We care more about your ability to learn and build than the number of technologies listed on your resume.

Strong programming fundamentals in Go and Python.

Exposure to Hands-on exposure to Kserve, Kubernetes, KubeRay, Ray Serve or distributed computing.

A good understanding of operating systems, Linux, networking and basic computer architecture.

Familiarity with containers, Docker or Kubernetes through coursework, projects or self-learning.

Genuine curiosity about GPUs, parallel computing, AI infrastructure or distributed systems.

Ability to read technical documentation, explore unfamiliar codebases and debug problems independently.

Evidence of building something beyond classroom assignments — GitHub projects, technical experiments, open-source contributions or personal prototypes.

Good to have — but not mandatory

Hands-on exposure to Kserve, Kubernetes, KubeRay, Ray Serve or distributed computing.

Experience deploying LLMs using vLLM, Triton or SGLang.

Familiarity with NVIDIA CUDA, GPU memory, VRAM utilization or GPU profiling.

Understanding of GPU MIG, time-slicing, device plugins or resource scheduling.

Experience with benchmarking, Linux performance tools or observability systems.

You do not need to know all these technologies before joining.

We are looking for strong fundamentals, demonstrated curiosity and the ability to learn quickly.

What you will gain

Hands-on experience building a real AI infrastructure product.

Exposure to GPU resource management, model serving and cloud-native systems.

The opportunity to experiment with leading open-source projects and contribute working code.

Mentorship from engineers building cloud and AI infrastructure.

The possibility of a full-time engineering opportunity based on performance and business requirements.

Who should apply?

Students in their pre-final or final year of Computer Science, IT, Electronics or related engineering disciplines, as well as recent graduates who enjoy systems programming and infrastructure engineering.

We welcome self-taught engineers and candidates from non-traditional backgrounds who can demonstrate their skills through practical projects.

The goal is not simply to complete an internship. The goal is to build engineers!!

If you’re looking for a role where thinking matters more than titles, we want to talk.

#How We Evaluate You

How you think, not where you worked

How you write, not what tools you list

How you reason through complexity

📩 Apply only if you enjoy deep products and hard problems.