Lead research and development of multimodal foundation models using sensor, video, and physiological signals.
Design self-supervised pretraining and sensor fusion approaches for heterogeneous time series and computer vision.
Validate model outputs against gold-standard references and publish research at top venues.
Tonal builds an intelligent strength training platform that uses multimodal sensor data to personalize fitness coaching. It is a growing remote-first team with a research-driven culture focused on machine learning, biomechanics, and product impact.
Run day-0 media launches, creating same-day demos, starter code, and social posts for new image, video, and audio models.
Publish cookbooks and guides to build inspirational examples that combine models and modalities.
Drive search and AEO visibility by writing content and promoting media benchmarks so developers discover OpenRouter first.
OpenRouter is the leading AI routing and infrastructure layer that enterprises use to access, manage, and optimize LLMs across providers. They are a fast-growing, fully remote team with a culture of autonomy and trust.
Build and scale cloud-native services and ML inference infrastructure for real-time AI voice detection.
Drive end-to-end ownership of complex technical initiatives from design to production.
Collaborate with engineering, research, product, and SRE to deliver high-reliability systems.
Pindrop is a cybersecurity company delivering continuous identity verification across voice, video, and digital interactions. It protects billions of high-risk interactions annually and is backed by top investors, with a culture of innovation and recognition.
Work with subject-matter experts to create synthetic tasks for training AI agents across a range of professional and technical domains.
Design synthetic task-generation methods that produce diverse, realistic and learnable tasks.
Build systems and tooling to mutate, validate and improve synthetic tasks.
They build reliable, fair and open infrastructure for AI training data. The company is backed by Y Combinator, used by startups and Fortune 500, and scaling quickly.
Design and execute red team operations against GenAI systems, LLM pipelines, and cloud environments, simulating real-world attacks.
Develop novel attack chains combining GenAI vulnerabilities with infrastructure and API weaknesses, and perform full-scope penetration tests.
Partner with SecOps to improve detections, tune response workflows, and translate findings into practical remediation.
Pindrop is a cybersecurity company delivering continuous identity verification across voice, video, and digital interactions. Backed by leading investors including Andreessen Horowitz, IVP, and CapitalG, it protects billions of high-risk interactions for enterprises and has been recognized by TIME as one of the 10 Most Influential Software Companies of 2026.
Build, train, and evaluate deep learning models on financial data to improve credit modeling and loss prediction.
Own projects end to end, from hypothesis to deployed models in Tilt's ML pipelines.
Experiment with representation learning and agentic AI tools to accelerate data science workflows.
Tilt builds mobile-first financial products and machine learning-powered credit models that look beyond credit scores to recognize real potential. They are a small, virtual-first team spread across 14 countries, focused on collaboration, fairness, and expanding access to credit.
Develop and test deep learning modeling approaches using transformers, embeddings, and neural networks.
Build frameworks that integrate into ML pipelines to improve credit modeling and loss prediction.
Use agentic development tools and design specialized agents to speed up data science and business workflows.
Tilt is a mobile-first financial technology company that uses machine learning and real-time financial signals to provide credit products beyond traditional credit scores. With millions of customers worldwide, Tilt is a virtual-first, collaborative team spanning 14 countries and 12 time zones.
Conduct research on reinforcement learning algorithms for multimodal models, including diffusion-based approaches and unified frameworks.
Design and build scalable RL infrastructure and reward modeling to support distributed training and improve stability.
Develop benchmarking protocols, curate multimodal datasets, and advance next-generation RL paradigms for real-world applications.
Tether is a fintech company pioneering digital finance solutions, including the stablecoin USDT, and developing energy, data, and education products. The company is a fast, lean global team of top talent working remotely across the world, with a culture focused on innovation and transparency.
Take on difficult AI problems from hypothesis to deployment, developing new algorithms when existing approaches fall short.
Design rigorous benchmarks, ablation studies, and statistical evaluations comparing methods with leading academic and industry approaches.
Work across areas including training-data attribution, model influence, interpretability, embeddings, and multimodal similarity.
Sureel AI builds infrastructure for AI rights, licensing, attribution, protection, and intelligence starting in music. Acquired by Warner Music Group, they operate as an independent company combining startup speed with global reach, with a culture of high standards and low ego.
Lead ML/DL research on NGS data, setting technical direction and building models for Women Health and Organ Health.
Work with bioinformatics pipelines, design experiments, and ensure rigorous validation.
Mentor data scientists and contribute to IP through patents and publications.
Natera is a global leader in cell-free DNA testing dedicated to oncology, women's health, and organ health. Their team consists of highly dedicated statisticians, geneticists, and other professionals from world-class institutions.
Work at the forefront of AI-powered drug discovery, advancing frontier research and end-to-end applications across discovery, design, and translation.
Collaborate with a cross-functional team of top-notch scientists, engineers, and drug hunters to develop scalable biomedical discovery solutions.
Apply expertise in machine learning, NLP, and computer vision to turn real-world data into a discovery engine for precision health.
Recursion is a clinical-stage TechBio company decoding biology to radically improve lives through AI-native drug discovery. The company fosters a vibrant, collaborative culture of frontier research and rapid iteration, uniting experts across science, engineering, and business.
Evaluate GPU kernel tasks for technical accuracy, realism, and reproducibility.
Identify memory allocation bugs, execution bottlenecks, and errors in parallel computing logic.
Provide clear, actionable feedback to improve AI training and evaluation workflows.
A partner company operating a specialized AI training project to improve the technical accuracy of AI systems. The company seeks experienced technical professionals for freelance, remote evaluation work, emphasizing precision and evidence-based assessment.
Train and fine-tune generative models for photorealistic and stylized images (text-to-image, image-to-image, video-to-image, IP-adapters).
Build models that understand visual context, including image classifiers, rankers, and image/video captioning.
Lead a team of three engineers: set priorities, review experiments, run 1:1s, mentor, and own the team’s results.
Social Discovery Group (SDG) is a group of social discovery companies that solves loneliness and disconnection by transforming virtual intimacy into the new normal. With an international team of digital nomads working remotely worldwide, SDG is a two-time “Great Place to Work” winner (USA & Japan, 2024–2025) and a Top-5 Company for Work-From-Anywhere Jobs (FlexJobs, 2025).
Develop and deploy machine learning models to detect web-based malware at scale.
Use production data to train and evaluate models for accurate threat detection.
Collaborate with security analysts to optimize detection and cleanup processes.
The company specializes in cybersecurity, using machine learning to detect web-based malware. It operates a remote-first culture with dedicated technical resources and a focus on collaboration and innovation.
Serve as the scientific authority on ground truth, testing AI agents on metabolomics data analysis.
Build tasks and datasets for benchmarks covering raw LC-MS, GC-MS, and NMR workflows.
Collaborate with a global team to push the boundaries of AI in biological reasoning.
Latch builds rigorous scientific benchmarks for AI agents across biology. The company works with frontier labs and pharma and fosters a culture of radical transparency and collaboration.
Develop novel hybrid machine learning and mechanistic models of bioprocess systems, from research proof-of-concept to production-ready use.
Own research projects end-to-end, collaborating with wet lab, dry lab, software teams, and leading academic and industrial partners.
Build transferable, uncertainty-aware models and real-time control layers for fermentation across organisms and scales.
Differential Bio is unlocking the potential of the bioeconomy by fixing how biological R&D and scale-up are done. They are an early-stage startup with a small but mighty team, focused on cutting-edge modelling and real-world bioprocess impact.
Explore complex datasets to identify fraud patterns, attack methods, and behavioral signals.
Build machine learning models for fraud detection across account opening, account takeover, and identity risk.
Write clean, well-tested code and work with engineering to bring models and features into production.
Experian is a global data and technology company operating across financial services, healthcare, automotive, and more. They have a team of 25,200 people in 32 countries and a people-first, inclusive culture.
Train, evaluate, and deploy spam-detection models, building and maintaining training corpora and analyzing precision and recall.
Develop production components of a new email-filtering engine in Go, including inference services and streaming message parsing.
Investigate customer-reported incidents and improve detection quality through rules, datasets, or model retraining.
This partner company develops advanced machine learning technology to protect email infrastructure against spam, phishing, and other evolving cyber threats. It operates a production-scale security solution filtering email across approximately 5,000 hosting servers and values remote-first, flexible working with a focus on engineering excellence.
Design, train, and ship speech-to-speech and real-time voice models to improve conversational quality on live calls.
Build and improve the production speech stack including neural codecs, streaming inference, and turn detection.
Diagnose real-world failures like telephony noise or overlapping speech and turn them into model, data, or system improvements.
Salient builds AI agents for regulated financial services. They are backed by a16z and Y Combinator, with $75M raised in Series A, cash flow positive and profitable, serving over 20% of the auto lending industry.
Build systems to detect PII, quasi-identifiers, credentials, and design transformations based on data type and downstream use case.
Develop and benchmark detection approaches combining rules, statistical models, classifiers, and LLM-based methods.
Create evaluation frameworks measuring privacy risk and retained data utility, including recall-weighted metrics and adversarial re-identification tests.
The client is a fast-growing AI company building reliable, fair and open infrastructure for AI training data. They are backed by Y Combinator, used by startups, Fortune 500 companies and frontier labs, and scaling quickly to meet demand.