Multimodal AI Systems Engineer

About Link Horizon Link Horizon is building a system that understands what’s happening on construction sites - in real time - and why it matters. We combine persistent, on-site aerial data capture with project schedules, contracts, and design files to create a continuously updating, structured view of a construction project. Our platform transforms daily drone flights into actionable insight, automating core workflows like progress tracking, site inspections, and contract administration for developers, contractors, and consultants. We operate at the intersection of hardware and software. On-site, we deploy and manage drone-in-a-box systems that capture high-frequency aerial data across large-scale land development projects. Off-site, we’re building the intelligence layer - a multimodal AI system that fuses vision data with project documentation to model site activity, track progress against plan, and surface what actually matters to project stakeholders. We’ve recently entered a design partnership with one of Ontario’s most innovative home developers and are launching our first pilot in Ottawa in May 2026. This is a critical moment: the product is being shaped in real time, directly on active construction sites. We’re a small, fast-moving team looking for people who want to build at the frontier of AI, robotics, and the built world to define a new category of site intelligence. About The Team Link Horizon is a small, engineering-first team moving quickly at an early stage. We work in tight loops, and will be deploying on real construction sites, learning from what actually happens, and shipping improvements continuously. We’re currently a two-person team and are now growing to support our first pilot deployment starting in May 2026. Both founders are University of Waterloo Mechatronics Engineering graduates, and our team actually originated from our final-year capstone project. Since September 2025, we’ve been backed by the University of Waterloo’s Velocity startup incubator and have been operating out of their Kitchener office up until now. We’re temporarily relocating to Ottawa from May-August 2026 for our first pilot and will be back in Kitchener in the Fall. We focus on building systems that work in the real world, on active sites, with imperfect data and real constraints. Engineers spend time in the field, decisions are made quickly, and ownership is high. This is a place for people who want to build from first principles and have a direct impact on the product and the company. The Role This is not a modeling role - it is a systems role. You will build the layer that takes everything Link Horizon collects and knows about a construction site and turns it into structured understanding: what is happening, what was supposed to happen, and what the gap means. The work spans document ingestion, CAD interpretation, computer vision, geospatial reasoning, data fusion, and multimodal AI. You will build pipelines that ingest project documents, interpret vector data from engineering drawings, run segmentation and detection models against drone imagery, and fuse what is observed on site with what was planned. All of it from the ground up. You will work directly with the founders to define system architecture, product direction, and how the platform evolves from first pilot deployment to production scale. You will also work closely with the broader team responsible for platform development and field data collection, with the opportunity to be trained to fly drones and visit sites regularly. What You'll Work On Build pipelines that ingest and interpret project documents, CAD files, schedules, and other unstructured inputs into structured, queryable representations Develop computer vision and segmentation models for object detection, state recognition, and change tracking on construction sites Build systems that fuse real-world site state (from drone data) with planned state (from drawings and schedules) to identify progress, deviations, and gaps Design temporal tracking systems that represent how a site evolves over time against planned intent Architect hybrid systems that combine deterministic data processing (geometry, schedules, structured data) with AI models (vision, LLMs, multimodal systems) Apply LLMs where appropriate for reasoning, interpretation, and interaction - not as a substitute for structured systems Own end-to-end development from prototype to production Coordinate with field operations to understand how data is collected and ensure your systems reflect real-world conditions Must-Have Experience with computer vision systems (object detection, segmentation, change detection) applied to real-world data Experience building multi-stage data pipelines that process unstructured or multimodal inputs into structured outputs Familiarity with geospatial concepts and working with spatially referenced data Strong proficiency in Python and backend system development Competency, familiarity, and willingness to experiment with and leverage cutting edge AI tools to accelerate everyday development workflows Comfort operating in a highly ambiguous, fast-moving environment Nice-to-Have Experience building systems that combine vision and language models to process real-world data Experience with photogrammetry, 3D reconstruction, or point cloud processing Experience working with CAD or other engineering design data formats Experience with real-time or streaming data systems Background in robotics, drones, or physical-world systems Familiarity with construction, surveying, or land development Compensation & Benefits $75,000–$150,000 base salary + bonus Strong health and benefits package Full autonomy to choose your tools, models, and development workflow (e.g., LLMs, coding environments, infrastructure) with budget provided High-ownership role with direct impact on product and technical direction Direct collaboration with founders in a low-bureaucracy environment Rapid iteration cycles with real-world deployment on active construction sites Daily team lunches and paid dinners when working late Additional Context This role will be permanently based out of Kitchener at the University of Waterloo’s Velocity startup incubator. During May-August 2026, you will be asked to temporarily relocate to the Ottawa area to support our initial pilot deployments with our design partner. Housing costs will be covered when temporarily relocating to Ottawa for the pilot project. Exceptional candidates who cannot temporarily relocate to Ottawa for May-August 2026 may be considered. You will also be trained and certified to operate drones to provide operational coverage when needed. How to Apply Interested applicants should submit their interest via the form at: https://apply.linkhorizon.com/ For any questions, please contact Tyler West at: [email protected]