Senior AI Engineer, Search

Stealth AI Startup; Stockholms kommun; On-site; SEK 650,000 - 900,000; Full time

Job summary

Stealth AI Startup in Sweden seeks a talented NLP/IR engineer to build production-grade semantic search systems, dense retrieval, and RAG pipelines. You will translate ideas into scalable architectures and deliver interpretable results for real users.

Candidates should have MSc/PhD in NLP/IR, hands-on experience with transformers and Hugging Face, and a track record of shipping systems used by people.

Qualifications

• MSc in NLP, Information Retrieval, Computational Linguistics, or Computer Science from a top university or equivalent standing
• Production experience building semantic search or information retrieval systems that real users depend on
• Architectural understanding of transformer models; deep comprehension beyond surface-level APIs
• Proficiency with the Hugging Face ecosystem, fine-tuning, evaluation, and production deployment

Responsibilities

• Build query understanding pipelines covering intent classification, named entity recognition, semantic parsing, and contextual query expansion across multilingual distributions
• Design and build dense retrieval infrastructure with bi-encoder and cross-encoder architectures for semantic search
• Build re-ranking systems to surface the most relevant results at query time
• Design and operate RAG pipelines that assemble information from multiple sources for a grounded generative interface
• Fine-tune pretrained language models on domain data to improve vocabulary coverage and pattern recognition
• Build precision evaluation infrastructure: query-level relevance, coverage, and longitudinal quality tracking

Job description

We are a well-funded, early-stage technology company on a mission to fundamentally redefine the way the world searches, discovers, and transacts, built on a philosophy that puts the user first at every step.

The user is not simply a means of transaction. They are a valued participant whose data works for them, not against them, enriching their world and their own experience within it, with them retaining full control.

Our mission is underpinned by proprietary, patent-pending technology built on a novel engine and system.

The Problem You Will Be Solving

The internet was built to inform the many, not to serve the individual. Platforms optimize for engagement, extraction, and consumption at global scale. AI has made this model even more efficient, with the user still largely remaining the input, not the full beneficiary.

The problem is not information. The world has never had more. The problem is that people give away a lifetime of data without receiving the personalized experience it should enable.

People should not have to search. They should simply find.

Who You Are

• Highly talented and motivated to build products with people, for people
• You thrive in zero-to-one environments and find genuine satisfaction in creating from scratch
• A fast-paced thinker with a relentless desire to learn, improve, and raise the standard around you
• Able to inspire, elevate, and collaborate; you make the people around you better
• A great communicator; you express complex ideas with clarity, precision, and confidence across all levels of the organization
• Able to transform ideas into working systems and technical milestones with light speed and precision
• Highly ambitious and action-driven; you thrive under pressure and find genuine energy in working at the edge of what is considered possible
• Capable of designing and building production-grade systems in a fourth of the time most experienced and “sane” professionals would consider possible

What You Will Do

• Build query understanding pipelines covering intent classification, named entity recognition, semantic parsing, and contextual query expansion across diverse and multilingual query distributions
• Design and build dense retrieval infrastructure including bi-encoder and cross-encoder architectures that retrieve and rank results based on deep semantic understanding of user intent
• Build re-ranking systems that refine candidate result sets to surface the most accurate and contextually relevant results at query time
• Design and operate RAG pipelines that assemble accurate, contextually appropriate information from multiple retrieval sources and deliver it through a generative interface grounded in current data
• Fine-tune pretrained language models on domain-specific data to enhance vocabulary coverage, entity understanding, and query pattern recognition where general-purpose models fall short
• Build precision evaluation infrastructure including query-level relevance assessment, coverage analysis, and longitudinal quality tracking across the full distribution of real user queries

Minimum Qualifications

• MSc in Natural Language Processing, Information Retrieval, Computational Linguistics, or Computer Science from KTH, Stockholm University, Chalmers, ETH Zurich, or an institution of equivalent standing
• Production experience building semantic search or information retrieval systems that real users depend on
• Architectural understanding of transformer models; not surface-level API familiarity but genuine comprehension of model behavior from first principles
• Proficiency across the Hugging Face ecosystem; fine-tuning methodology, evaluation, and efficient production deployment

Preferred Qualifications

• PhD in NLP, Information Retrieval, or a directly related field; given significant weight in our assessment
• Research publications in information retrieval, NLP, or related areas
• Experience designing retrieval systems that account for rich contextual signals beyond the query alone
• 2–6 years of industry or research experience

What We Offer

Expect the most talented, rigorous, and driven colleagues you have worked with; people who lift each other up and hold each other to the highest standard. This company is driven by ideas and innovation, regardless of age, position, or background. The best ideas must win and the people behind them recognized and rewarded.

A home for exceptional people and geniuses within their domain; those who have always known what they are capable of and have been waiting for the environment to prove it to the world, whilst first being seen and supported by their coworkers.

Compensation includes an attractive salary and meaningful equity ownership for an early-stage startup.

We are an equal opportunity employer committed to building an inclusive, high-performance culture where everyone can do their best work and reach their full potential. We welcome applicants of all backgrounds and do not discriminate on the basis of race, color, religion, national origin, gender, gender identity or expression, sexual orientation, age, disability, or any other characteristic protected by applicable law.

Application

Please note that all applications and interviews are conducted in English. Fluency in written and spoken English is a requirement for all roles.

All applications are treated with strict confidentiality.