Senior Machine Learning Engineer - AI/Agentic System

Arbisoft; Lahore; Hybrid; PKR 3,200,000 - 6,000,000; Full time

Job summary

Arbisoft is seeking a Machine Learning Engineer – AI/Agentic Systems to join our team. You will work on globally scaled travel-related AI projects, handling terabyte-scale datasets and building labeled data pipelines.

Ideal candidates have strong Python and PyTorch skills, experience taking ML projects from concept to production, and familiarity with Docker and Kubernetes for deployment. Hybrid work with 3 days in-office is required.

Qualifications

• 5+ years of experience in machine learning engineering or similar roles.
• Proficiency with Python and PyTorch.
• Experience owning ML projects end-to-end from problem to production.
• Experience working with large-scale datasets (text and/or image).
• Exposure to LLMs and agentic AI concepts is a plus, with practical experimentation encouraged.

Responsibilities

• Work with very large terabyte-scale datasets across text and image domains.
• Build labeled datasets using manual labeling, statistical methods, and LLM-based approaches.
• Own the full project lifecycle from R&D to production deployment.
• Solve open-ended ML/AI problems, including data processing and model deployment.
• Develop and maintain data pipelines to support scale and performance.

Skills

Python

PyTorch

End-to-end ML

Large-scale datasets

LLMs & agentic AI

Data pipelines

Tools

Docker

Kubernetes

Job description

Experience: 5 years +

Work Mode: Hybrid, 3 days/week from office (mandatory)

Base Location: Islamabad, Lahore

Data Scale: Terabyte-scale datasets, across text and image domains

Core Domains: Text, Image, and Agentic AI

Core Stack: Python, PyTorch, Agentic AI/LLMs, Docker, Kubernetes

Arbisoft is hiring a Machine Learning Engineer – AI/Agentic Systems for one of our client projects.

Our client is a globally recognized travel search platform, used by millions of travelers worldwide to search, compare, and book flights, hotels, and rental cars. The client is investing heavily in AI-powered and agentic travel planning experiences, alongside its core large-scale search and personalization systems.

What You'll Be Doing

• Working with Large-Scale Data: We work with very large datasets, often at terabyte scale, across both text and image domains.
• Building Labeled Data: Many of our problems text, image, and agentic AI start with little or no labeled data. You'll get time and ownership to build/curate labeled datasets using manual labeling, statistical/ML techniques, and LLM-based approaches.
• Owning the Full Project Lifecycle: Every team member is responsible for the complete project lifecycle, from R&D and experimentation to production deployment.
• Solving Open-Ended Problems: We work on image, text, and agentic AI problems, so experience solving open-ended ML/AI problems is important.
• Engineering Data at Scale: You should be comfortable handling, loading, processing, and parallelizing large-scale datasets efficiently.

What We're Looking For

• Strong hands-on experience with Python and PyTorch
• Demonstrated experience owning ML projects end-to-end from problem definition to production deployment
• Experience working with large-scale datasets (text and/or image) with strong data engineering fundamentals
• Exposure to building labeled datasets from limited/no ground truth via manual labeling, statistical/ML techniques, or LLM-based approaches
• Familiarity with LLMs and agentic AI concepts production experience is a strong plus, but hands-on experimentation, side projects, or research exposure will also be considered
• Working knowledge of Docker and Kubernetes for deployment
• Ability to independently explore a problem, experiment with different approaches, build the required data pipeline, and take the solution to production

Nice to Have

• Experience with distributed data processing tools (Spark, Ray, Dask)
• Open-source contributions or public technical writing related to ML/LLMs
• Prior experience in a product company or applied AI/data-focused software house serving global clients

What We Value

Someone who can independently explore a problem, experiment with different approaches, build the required data pipeline, and ultimately take the solution to production.