AL/MI Engineer

About the Company Zuma Labs is a disruptive technology startup building force multipliers for systemic industries. Our first two products, Venetian and Squawk, target the over-the-counter (OTC) trading markets. We work closely with our customers and count a large, publicly traded brokerage among our partners. Trading technology is an old industry… and we’re here to change that. Like many startups, we value creativity, entrepreneurship, and real passion for problem-solving. But what we’re really looking for are born-and-bred hackers, people with a healthy disrespect for the status quo and the drive to build something better. About the Role As an AI / ML Engineer, you will work directly with the founder, engineers, brokers, and traders to design and deploy intelligent systems that have immediate real-world impact. You’ll be given broad, high-level objectives aligned with the company vision - and significant autonomy in how you execute them. We prioritise working solutions over perfect answers, rapid iteration over theory, and measurable impact over vanity metrics. We’re looking for a driven self-starter who takes confidence in their work, yet isn’t afraid to ask questions, challenge assumptions, or seek advice. Nobody has all the answers - and the strongest engineers are those who learn fastest. Responsibilities * Develop and deploy LLM-powered systems using Torch and SGLang. * Build, fine-tune, and optimise embedding models and reranking models. * Implement parameter-efficient fine-tuning pipelines (PEFT / LoRA adapters). * Design and execute data curation strategies, including synthetic data generation for training and evaluation. * Work closely with product and domain experts to translate real trading workflows into ML solutions. * Run experiments, evaluate model performance rigorously, and ship improvements quickly. * Contribute to infrastructure supporting training, evaluation, and production deployment. Qualifications * Strong hands-on experience with PyTorch (Torch). * Experience working with LLMs, embeddings, and reranking models. * Practical experience with PEFT techniques, including LoRA adapters. * Experience with dataset construction, cleaning, and synthetic data generation. * Strong understanding of ML fundamentals (evaluation metrics, optimisation, generalisation, etc.). * General understanding of supporting infrastructure: git, containers, microservices, etc. * Comfortable working with numerical data and analytical problems (finance/trading knowledge not required). Preferred Skills * Experience deploying ML systems into production environments. * Familiarity with RAG architectures, prompt engineering, or hybrid retrieval systems. * Experience building low-latency or real-time ML systems. * Exposure to experimentation frameworks and model benchmarking. * Cloud infrastructure and containerised ML workflows. * Experience working in a startup or high-autonomy product environment. Build AI systems that power real users in demanding, systemic industries. Operate in a high-impact startup environment where your work directly shapes the product. Collaborate with founders, engineers, and traders in a fast-moving, intellectually rigorous team. Work at the intersection of LLMs, real-world markets, and applied engineering - not research for research’s sake. Equal Opportunity Statement Zuma Labs is committed to diversity and inclusivity in the workplace.