Job Title

About the RoleThe AI Engineer will be responsible for designing and building proprietary machine-learning models that reason about electrical and mechanical engineering work for use in automated scoring and evaluation.ResponsibilitiesDesign and build proprietary machine-learning models that reason about electrical and mechanical engineering work (schematics, PCB layouts, CAD geometry, simulations, and design workflows) for use in automated scoring and evaluationDevelop novel model architectures and training pipelines for technical reasoning, not just text generation — including multi-modal reasoning over CAD/ECAD artifacts, simulation outputs, and candidate interaction tracesTranslate real engineering tasks (circuit design, debugging, system integration, mechanical design tradeoffs, etc.) into machine-interpretable representations that models can evaluate reliably and deterministicallyBuild and operate the full learning loop for these models: (a) data generation from real assessment executions (b) trajectory capture (tool use, intermediate designs, decisions) (c) failure analysis (d) targeted dataset curation and “golden” supervision (e) continuous evaluation and model iterationCreate scoring systems that are robust, defensible, and hard to replicate, forming the technical foundation of assessment platformWork closely with product and platform teams to deploy these models into production scoring pipelines used by real hiring decisions RequirementsStrong background in machine learning, including deep learning and modern foundation model architectures.Experience designing and operating end-to-end training and evaluation pipelines for production ML systems.Practical experience with retrieval-augmented generation (RAG) systems and vector databases for large-scale knowledge and artifact retrieval.Experience working with noisy, real-world labeled datasets, including data cleaning, schema design, and quality control.Hands-on experience with reinforcement learning, including one or more of: (a) reinforcement learning from human feedback (RLHF) (b) preference modeling and reward model training (c) policy optimization for multi-step or tool-using agentsExperience building or training models on multimodal data (text, images, video, or structured technical artifacts such as diagrams or CAD files)