MACHINE LEARNING RESEARCH INTERN

About Joko helps consumers shop smarter by empowering them to find what they need, make informed decisions, and save money. Founded in Paris, Joko is a tech company and certified B Corp with an international team of 80+ talents spread across Paris, Barcelona, and New York (and beyond). More than 5 million users already save money every day at over 10,000 merchants. From cashback and automatic coupons to price alerts and carbon tracking, we continuously expand our products to make shopping smarter. We are now building an AI‑powered shopping assistant that helps users find the best products by price, quality and environmental impact. Having reached profitability in our core market, we are scaling globally with a strong focus on the US. Job Summary At Joko, we help consumers shop smarter. Our mission is to revolutionize shopping, empowering people to find what they need, make informed decisions, and save money. This position will be located in Paris. Research at Joko Our research team tackles two main challenges: Building the universal product catalog: We are developing the world’s largest universal product catalog, unifying products sold across e‑commerce worldwide. This requires understanding any web page to extract structured information and clean and standardise data from millions of heterogeneous sources in near real time. Developing the AI Shopping copilot: On top of the catalog, we are building an AI copilot that helps users navigate our product universe through natural conversation. This goes beyond traditional RAG systems, combining full‑text and vector search across massive datasets and leveraging LLMs to ensure a reliable, fast, and cost‑efficient experience. We rely on LLM‑based methods and have developed state‑of‑the‑art approaches centered on fine‑tuning relatively small models. While much research remains to optimise resource usage and scale our systems, we explore new architectures that push the boundaries of what LLMs can achieve in large‑scale e‑commerce. What you’ll do As a Machine Learning Research Intern, you will work closely with our engineering teams, owning a project that can be deployed to production and monitored for impact on hundreds of thousands of users. You will conduct experiments, literature reviews and theoretical developments, and choose your research direction based on promising results. Improving the Universal Catalog Pipeline Enhancing the performance and scalability of our LLM‑based data processing pipeline Exploring fine‑tuning for specialized tasks and techniques to reduce model size (quantisation, pruning, distillation) Developing rigorous evaluation methods, notably using LLMs as judges Improving Search in the Catalog Benchmarking search techniques that combine full‑text and vector search Exploring the use of images to enhance search performance Identifying and evaluating the most effective LLM‑based embedding methods Improving the AI Copilot Benchmarking and fine‑tuning models to improve both performance and latency Working on model optimisation (size reduction, distillation, quantisation) Establishing rigorous evaluation pipelines Your Responsibilities Research: Formalise objectives, conduct literature reviews, design new algorithms, and collect/transform data for experiments. Exploration & Ownership: Orient your internship toward research directions you deem most valuable to users. Implementation, Deployment & Monitoring: With support from the engineering team, integrate robust algorithms into production and monitor their real‑world impact. Perks Flexible remote (for fully‑remote positions): If you live in Paris, you can work from our office or from home with no constraints. If you live elsewhere, we provide access to a coworking space and a coworking budget. Work from anywhere (full‑time only): Full‑time employees can work from most countries for up to 3 months per year. After 4 years, spend one month in our New York office. Yearly offsite in amazing locations, plus a team‑building budget and quarterly in‑person gatherings. Contribution to your ClassPass subscription. 8‑week paid parental leave. Preferred Experience Technical & Problem‑Solving You have solid programming skills, with experience in Python and Machine Learning frameworks, and you are eager to keep improving. As a problem solver, you combine strong analytical skills, creativity, and a taste for tackling complex challenges. You are also tech‑savvy, constantly curious about emerging technologies. Mindset & Communication You have an entrepreneurial mindset: you like challenges, welcome feedback, and strive for excellence. You have strong communication skills, both written and verbal, and can explain complex topics in simple words. You are fluent in English (both spoken and written); mastering French is not required. Adaptability & Efficiency You are a fast learner, comfortable in any technical environment, and quick to adopt new technologies and practices. You value efficiency, make good use of productivity tools, and deliver on time even in projects with many stakeholders. You have great attention to detail. Preferred Background Education: You are a graduate student in a Machine Learning master from an engineering school, and looking for an end‑of‑studies internship. Experience: You have experience with research, from past internships or projects. Recruitment Process Intro call: Quick screening with the Hiring Manager or a recruiter. Step 1 – Team interview (45 min): Conversation with two Joko team members. Step 2 – Role‑specific assessments. For non‑tech roles: Take‑home case study followed by a 45 min debrief. For tech roles: Live technical interviews (Coding interview + System design interview). For research internships: An additional round may involve analysing and presenting a research paper. Step 3 – Founders interview (45 min): Conversation with two of our founders. References: Up to five calls with former colleagues or managers. ☕ You may also be invited for coffee with team members to get a feel for our culture. Additional Information Contract Type: Internship Location: Paris Possible partial remote #J-18808-Ljbffr