AI Research Engineer

Wisdomlabs | Artificial Intelligence

59m ago

On-site

Full Time

United States of America

Company Overview:

Everyone builds up knowledge and experience that could help someone else, but very little of it ever leaves their head. When they’re busy, asleep or just not in the room, the people who need it are left waiting.

TwinCafe, by Viven, gives everyone a digital twin to close that gap. A twin carries a person’s expertise and way of speaking, so it can talk with others for them, even about questions they’ve never been asked before. It draws only on what its owner has taught it, and when a question falls outside that, it hands the conversation back to them.

Educators, lawyers, coaches, recruiters, job seekers and creators use twins to stay reachable for students, clients and audiences, on the web, on mobile and on their own sites. Owners decide what their twin knows and can change it whenever they like.

Most AI products help people get more done. We’re building one that helps people share who they are.

Role Overview

We're seeking an AI Research Engineer to tackle TwinCafe's hardest technical problems, from the first idea to shipped product. This is a 0-to-1 role: you'll proactively identify opportunities to improve AI quality, define how to measure success, experiment with new approaches, and build the foundations to bring them into production. We're looking for engineers with strong foundations in the details of applied AI research, with a strong problem-solving and experimental mindset, and an ownership- and product-oriented mindset.

What You'll Work On

At TwinCafe, we're building digital twins that can deeply understand people, faithfully represent how they think, and create entirely new ways for humans to interact with AI and each other. You'll tackle problems across:

Agent Systems & Evaluations: How do we build agents that reliably reason and act on behalf of a specific person across different channels and use cases? How do we evaluate their behavior in open-ended scenarios without clear ground-truth answers?
Memory & Personalization: How do we build AI that genuinely understands someone and continuously learns? We work across pairwise and multi-agent memory, knowledge ingestion and retrieval, context management, and continual learning.
AI-Native Consumer Experiences: How do we make creating, teaching, and interacting with digital twins feel natural and intuitive? We're exploring conversational onboarding, generative UI, low-latency voice agents, interactive avatars, personalized discovery, intelligent simulations, and more.

Must-Haves

Applied AI & Agent Systems: You’ve built and improved AI systems that real people use, and you understand how modern LLMs and agents work, including context engineering, tool use, retrieval, and memory. You have a feel for what today’s models do well and where they fall short.
Evaluation-Driven Development: You measure before you trust. You’ve designed evaluations, such as datasets, synthetic data, LLM-based graders, or careful failure analysis, and used them to improve quality, reliability, latency, or cost.
Research Depth & Technical Curiosity: You have a solid grasp of ML and LLM fundamentals, follow new research, and enjoy turning promising ideas into working systems, often by combining new approaches with proven ones.
Nice-to-Haves
Personalization & Continual Learning: Experience with long-term and pairwise memory, multi-agent memory systems, user modeling, knowledge extraction, personalized agents, or systems that continuously learn and adapt from interactions.
Multimodal & Real-Time AI: Experience building low-latency voice agents, speech systems, interactive avatars, or multimodal AI experiences.
Model Training & Self-Improvement: Hands-on experience with LLM fine-tuning, reinforcement learning, preference optimization, reward modeling, synthetic data pipelines, or self-improving agent systems. Familiarity with PyTorch or frameworks such as TRL, Tinker, or Prime Intellect.
Search & Recommendation Systems: Experience with semantic retrieval, search and feed ranking, recommendation systems, or large-scale information retrieval.
Research & Open Source: Contributions to AI research, open-source projects, or novel implementations of research papers.
Startup Experience: Experience at early-stage startups, building new products from scratch, or leading ambitious technical projects with high autonomy.
Engineering & Product Ownership: Strong software engineering skills and a 0-to-1 mindset. Obsesses over real user experiences, proactively identifies problems worth solving, and owns solutions from research and prototyping through production. Comfortable working across AI, backend, data infrastructure, and product.

We don't expect you to have experience in every area. We care most about deep technical expertise in one or more domains, strong research instincts, and the ability to learn quickly and solve problems across the AI stack.

Equal Opportunity Employment:

TwinCafe is committed to equal employment opportunity and non-discrimination for all employees and qualified applicants. We value diversity and are dedicated to fostering an inclusive environment for all employees, regardless of race, color, sex, gender identity or expression, age, religion, national origin, ancestry, citizenship, disability, military or veteran status, genetic information, sexual orientation, marital status, or any other characteristic protected under applicable law.

If you are passionate about AI and eager to make an impact, we would love to hear from you. Apply today to join the TwinCafe team and help us build the future of AI.

Required Languages

🇬🇧 English