AI Infrastructure Intern — Build Production LLM Stack

Grayhat Developers Private Limited; Islamabad; On-site; PKR 446,000 - 670,000; Full time

Benefits offered by this job

Paid internship

Mentorship from engineers

Portfolio-building experience

Job summary

Grayhat Developers Private Limited is seeking a curious intern to join our AI infrastructure team. You will work directly on our self-hosted LLM stack, addressing real production challenges in model performance and usability.

You will build tooling, explore quantization and other optimization methods, and collaborate with the broader engineering team to shape how we deploy AI internally. This is an onsite role with hands-on learning and mentorship.

Qualifications

• Solid Python skills and comfort with Linux/server environments.
• Solid understanding of how transformer-based LLMs work, not just prompting.
• Exposure to LLM serving frameworks like Ollama, vLLM, llama.cpp, or TGI.
• Genuine interest in model optimization and quantization techniques.

Responsibilities

• Dig into bottlenecks in our self-hosted LLM stack around latency and hardware utilization.
• Build internal tools and interfaces for non-technical users to interact with the AI stack.
• Experiment with model optimization techniques like quantization, pruning, and batching within hardware constraints.
• Identify gaps in tooling and create lightweight integrations or wrappers to close them.
• Maintain clear engineering notes so work is reproducible for the next engineer.

Skills

Python

Linux/server

Transformer-based LLMs

Builder mindset

Education

CS/AI/Data Science student or fresh graduate

Tools

Ollama

vLLM

llama.cpp

TGI

Job description

Grayhat Developers Private Limited is seeking a curious intern to join our AI infrastructure team. You will work directly on our self-hosted LLM stack, addressing real production challenges in model performance and usability.

You will build tooling, explore quantization and other optimization methods, and collaborate with the broader engineering team to shape how we deploy AI internally. This is an onsite role with hands-on learning and mentorship.