ML Engineer

Machine Learning Engineer @ Visia New York, NY (Hybrid/In-person) About Us Visia is the first multimodal AI platform custom built for heavy industry. Visia’s full-stack physical intelligence platform includes robust sensing systems across imaging modes (cameras, X-rays, cargo X-rays, LiDAR), foundation vision-language models that convert raw sensor data into structured operational intelligence, and software-driven Field Engineering that drives real transformation on-site at some of the world’s largest industrial operations. We deploy our hardware systems powered by our model flywheel across recycling facilities, steel mills, aluminum smelters, ports, and waste-to-energy plants, processing over 1 billion data points a year.

When customers adopt Visia they aren’t buying a point solution – they are parting with a platform business that turns unstructured optical data into real-time, actionable intelligence for their operations.

What We'Re Looking For

We're looking for a Platform Engineer to own and accelerate Visia's data flywheel — the engine that makes our fine-tuned models smarter with every deployment. You'll primarily build the tooling and automation that powers our annotation, model training, and data pipelines, ensuring that every image we process makes Solstice better at understanding the physical world. This is a high-leverage role.

The systems you build directly determine how fast we can onboard new customers, new sensor modalities, and new industry verticals. You'll also pitch in on custom Field Engineering implementations when needed and contribute to internal agentic tooling that helps the team move faster.

Responsibilities

Data Flywheel & Model Tooling Automate and improve our instance-segmentation and vision-language model annotation tooling, creating smooth, efficient workflows for our annotators and model managers Build and maintain the pipelines that move data from deployed sensors → annotation → model training → production deployment Develop tooling for model performance monitoring, error analysis, and continuous improvement across customer sites Work closely with ML engineers to accelerate the feedback loop between deployed models and training infrastructure Field Engineering Support Contribute to custom Visia implementations for enterprise customers — building integrations with scale systems, ERPs, PLCs, and customer-specific reporting Support Forward Deployed Engineering (FDE) engagements by building reusable tooling and templates that make each deployment faster than the last Help translate customer operational needs into technical specifications and working softwareInternal Tooling & Automation Build agentic tooling to accelerate GTM workflows — automated sales deck generation, customer research, and outreach Collaborate with the CEO on automating operational workflows Contribute to Visia's overall product roadmap and help shape the future of physical intelligence in heavy industry

Who You Are

You have strong programming skills in one or more commonly used programming languages (we use Python, TypeScript, and Go) You move fast, ship iteratively, and care about building things that actually work in production You're comfortable working with at least some of the following and eager to learn the rest: Python backend development (Flask, FastAPI) Frontend development (React / React Native) Cloud computing (GCP, AWS, or Azure) CI/CD, IaC, and containerization (Docker, Terraform, GitHub Actions) Databases (SQL, dbt, Postgres, ClickHouse) Bonus: experience with computer vision pipelines, annotation tooling, or ML infrastructure You have proof of your engineering ability through GitHub projects or work experience What You Might Work On Building automated annotation pipelines that use SAM and vision-language models to pre-label datasets, cutting model iteration cycles by 5-10x Creating model performance dashboards that surface accuracy regressions across customer sites before they become customer issues Integrating