This is a job posting for a Machine Learning Engineer at Hestus, a YC-funded startup building AI-powered CAD tools for hardware development. They are looking for someone to develop and deploy custom ML models and prototype tools at the intersection of mechanical engineering and AI. The role is full-time, onsite, with competitive salary and equity.
What they want, where you stand, and the exact résumé edits to qualify.
Biggest lever: Develop and deploy a custom ML model (e.g., a small image processing or regression model) and integrate it into a web application to demonstrate end-to-end ML system building.
A starter prompt for Claude Code, what you'll need, and how to reach them.
You are a senior Machine Learning Engineer. Your task is to design a high-level technical architecture for an AI-powered CAD tool, specifically focusing on an MVP that automates a common, time-consuming task in mechanical engineering design, such as generating initial design options based on functional requirements or optimizing a component for weight/strength. The MVP should be a web-based tool. Use Python for the ML backend and a modern web framework for the frontend. Specify the core ML model type and required data sources. Outline the key components, data flow, and deployment strategy using modern cloud services. Key requirements: - **Problem Focus**: Automated conceptual design generation for simple mechanical parts (e.g., brackets, levers) based on textual or parametric inputs. - **ML Approach**: Propose a suitable ML model (e.g., generative adversarial networks, variational autoencoders, or reinforcement learning) for 3D geometry generation or optimization. - **Data Strategy**: Identify types of training data needed (e.g., existing CAD models, simulation results, engineering specifications) and strategies for data acquisition/synthesis. - **Architecture**: Design a distributed system with a clear separation between frontend (user interaction), backend (API gateway, business logic), and ML inference/training services. - **Tech Stack (Backend)**: Python, FastAPI for API, PyTorch/TensorFlow for ML, Dask/Ray for distributed computation. - **Tech Stack (Frontend)**: Next.js 16 App Router, React 19, Tailwind v4 for UI, Three.js for 3D visualization. - **Database**: Neon Postgres for metadata and user preferences. - **Deployment**: Vercel for frontend/API, AWS/GCP for ML inference/training microservices. - **MVP Scope**: Focus on generating one type of part (e.g., a simple bracket) with basic parameter inputs and showing a 3D preview. Users should be able to input constraints (e.g., max load, material) and get suggested designs. Provide the architecture diagram components, API endpoints, and a high-level implementation plan for the MVP.
Hestus | Machine Learning Engineer (Python, ML) | Peninsula (Onsite) | Full-time | $130k–$190k + equity | YC-funded We're building AI-powered CAD tools to transform hardware development — think next-gen design workflows powered by custom ML models. We're a fast-moving startup, working with bleeding-edge tech at the intersection of mechanical engineering and AI. We're looking for a versatile ML engineer to help us scale: develop and deploy custom models and prototype tools that make complex design work simpler and faster. https://www.hestus.co/careers for more information
Focus on a practical ML project. Pick a small, publicly available CAD-related dataset (e.g., simple geometric shapes, material properties) or a general image dataset. Learn a specific ML framework (e.g., PyTorch or TensorFlow) to train and evaluate a custom model. Then, deploy this model as a microservice (e.g., using FastAPI + Docker) and integrate it into a simple web front-end to demonstrate the full lifecycle of ML model development and deployment (approx. 2-3 months). This would also provide a stepping stone to understanding ML in design tools.
New: deploying and managing ML inference services on cloud infrastructure, specifically for Python ML models.
Learn it: AWS docs ↗
Get set up: Create an AWS account → an IAM user with least-privilege keys → configure the AWS CLI.
New: deep learning frameworks and model development for generative tasks (~1-2 weeks study).
Learn it: Search getting-started ↗
New: fundamental concepts of mechanical design, 3D modeling, and CAD workflows (~a few days).
Learn it: Search getting-started ↗
New: 3D rendering in the browser for visualizing CAD models (~1 week).
Learn it: Search getting-started ↗
Apply directly via the company careers page: https://www.hestus.co/careers
“This is a direct job application. Highlight relevant ML engineering experience, particularly any work with 3D data, generative models, or high-performance computing, and express enthusiasm for the intersection of AI and mechanical engineering.”
Open the original ↗