GNN Developer

The company is changing the way people design and manufacture houses, cars, and satellites by empowering designers and engineers to automate technically accurate 3D models that can be manufactured. The team, composed of engineers working aggressively and passionately, is focused on revolutionizing the traditional CAD design, engineering, and manufacturing process. Their goal is to enable people building the future to do so faster.

Role Overview

You will be working directly on all our GNN efforts such as building architectures & modules, working on graph structures extensively to identify, represent & modify various elements, properties, hierarchies etc. You will be a part of the core GNN team & working closely with other teams to Augment Human Capabilities.

If you resonate with the mission & are passionate about solving the hard problems relentlessly, aren’t afraid to try the bold ideas & fail (most of the times), this is the perfect role.

Responsibilities

• Develop and apply advanced graph neural network models to identify, verify & modify complex structural assemblies and component representations of detailed 3D models including but not limited to apartments, satellites, buildings and cars.
• Develop graph data structures for identification, hierarchical representation, modification and verification of 3D model meshes, topology etc.
• Understand and improve the effect of computational graph representation on the model execution, performance on different platforms, GPU, CPU etc.
• Experiment & apply strategies for continuous learning in the context of 3D data, enabling AI systems to adapt and improve over time
• Research and experiment with emerging technologies and tools in the graph and AI domains, ensuring Augrade’s solution stays ahead of the curve

Requirements

• 3+ yrs experience OR some really cool (technically) projects.
• Strong expertise in GNN frameworks and architectures.
• Strong understanding of 3D environments.
• Strong skills in data & database management.
• Experience with knowledge graph stores such as Neo4j (or its equivalent TigerGraph, Neptune, etc), and surrounding semantic technology (OWL, RDF, SWRL, SPARQL, JSON-LD) is a significant plus
• Ability to work comfortably with computations graphs and IRs (compiler background a strong bonus)
• Deep understanding of modern vector/embedding-based graph representation techniques with proven experience in designing and implementing algorithms for their generation
• Enjoy collaborating within and across teams
• Fast learner.
• Must be amenable to work 40+ hours in a week

Nice-to-haves

• Experience with LLMs, convex optimization, and physics-based simulation.
• Robust knowledge and experience in implementing graph analytics solutions in distributed systems, including community detection/understanding, pattern matching, sub-graph extraction, and importance ranking
• Experience in Architecture, Aerospace, or Automotive industries

Benefits

• Generous PTO / sick leave
• ESOPs ( For select few candidates )
• Visa sponsorship ( For select few candidates )
• And more as the team grows