• Define the core abstractions: Establish the shared schema of traits, physical fields and world states that every module reads and writes, a common contract for operators and evaluators, and the representation of EDEN’s procedural graphs. These graphs are design processes our teams author from real projects, which AI agents adapt to new sites, the decision engine searches and users inspect; their control nodes hold the executable programs that agents write and optimizers tune.
• Architect the search, optimization and agentic control loop: EDEN’s decision engine pairs AI agents that adapt graphs and write design programs with numerical optimizers. Design the harness that translates user intent into formal problem bounds, with guardrails that keep agent and optimizer output within expert-validated physical and ecological limits, while orchestrating parallel search, inner-loop tuning, evaluation budgets and diagnostic feedback.
• Design and build the runtime: Orchestrate physical simulation, learned surrogate models and long-running optimization at scale, with deterministic execution, content-aware caching, sandboxed code execution and deliberate trade-offs between fidelity and speed.
• Unify the prototypes: Bring today’s components and applications, built across different research and design environments, onto a portable, scalable foundation so new scientific modules and application verticals plug in cleanly without custom glue.
• Make validation part of the architecture: Ensure every model carries provenance, calibrated uncertainty and a documented range of conditions in which it holds. Make determinism, out-of-sample testing and physics-grounded checks on optimized designs platform features rather than one-off scripts.
• Design the data and delivery platform: Structure geospatial and ecological data ingestion, versioned site and run state, and the backend services and APIs that power EDEN’s interactive web interface and external design-tool integrations.
• Set engineering practice: Establish testing, reproducibility, CI and a clear line between research code and production code.

Key Goals & Outcomes

• First 90 Days:
• Map the existing components and applications, documenting what each one does, what it depends on and how it runs end-to-end.
• Write the architecture specification and implement the first versions of the shared data contracts and procedural graph runtime, reviewing them with the science and application teams.
• Unify the search and evaluation harness so agentic search, numerical optimization and simulation run through a single, cached lifecycle.
• Take one application end-to-end through the new architecture, from user request to validated result.
• A single, cohesive system in which new scientific modules and application verticals plug in cleanly without custom glue.
• Validation built into the platform, with provenance, calibrated uncertainty, determinism and physics-grounded checks as standard features.
• Clear engineering practices for testing, reproducibility, CI and the line between research and production code.
• An architecture that lets EDEN grow from one validated application to many.

Required Experience

• Substantial experience designing and owning complex compute- or data-intensive systems, ideally in simulation, scientific computing, production ML, mathematical optimization or geospatial platforms.
• A track record of turning research prototypes into reliable, well-structured products.
• Strong systems design and Python expertise, plus comfort with GPU workloads, parallel orchestration and cloud infrastructure.

Preferred Experience

• Experience with black-box, Bayesian or multi-objective optimization, quality-diversity search or reinforcement learning.
• Experience orchestrating and sandboxing LLM agents or program synthesis in production.
• Experience with surrogate modeling, differentiable simulation or GPU compute frameworks.
• Experience with procedural, DAG or node-based execution systems.
• Experience integrating with 3D web visualization, CAD/BIM or GIS software.
• An interest in ecology, earth observation or environmental modeling. This isn’t required, but the role is surrounded by it.

Technical Skills

• Systems architecture and design for compute- and data-intensive platforms.
• Expert-level Python.
• GPU workloads, parallel orchestration and cloud infrastructure.
• Runtime design for simulation and long-running optimization, including deterministic execution, content-aware caching and sandboxed code execution.
• Data platform design, including geospatial and ecological data ingestion, versioned state, backend services and APIs.
• Validation and reproducibility engineering, including provenance, calibrated uncertainty, determinism and out-of-sample testing.
• Engineering practice, including testing, CI and separating research code from production code.

Essential Qualities

• Real care for numerical reproducibility and whether a system’s outputs can be trusted, not only whether it runs.
• Clear writer and communicator for an audience of scientists, designers and engineers.
• Hands-on builder who writes the first versions of critical pieces personally, not only the specifications.
• Sound judgment on foundational decisions that are expensive to change later.
• Comfortable working across disciplines and teams, from platform engineering to scientific modeling to applications.