How to apply for Founding Software Engineer , Principal

About OnPeak Intelligence

OnPeak Intelligence builds weather decision intelligence for high-consequence industries, turning frontier atmospheric science into operational products that protect utilities and communities from wildfire risk. As a early-stage startup, it offers the rare chance to define the technical foundation of a company whose work directly impacts safety and infrastructure resilience.

About the role

This is a true 0→1 architect role: you will own the data model, ontology, and end-to-end platform that transforms continental NWP data, fire-spread simulations, and ML downscaling into reliable, scalable products for utility customers. Your architectural decisions in the first months will become the substrate the company runs on for years, making this one of the most impactful individual-contributor roles in the company.

A typical day

You might start your day reviewing the latest NWP data ingestion pipeline, then dive into designing a versioned ontology for utility assets that integrates with weather forecasts. Later, you'll pair with a scientist to optimize a GPU-based downscaling workflow, and end the day documenting architectural decisions that will guide the team for months to come.

Who OnPeak Intelligence is looking for

• Proven track record as a hands-on technical architect who has designed and built data-intensive platforms from scratch, ideally in weather, climate, geospatial, or scientific computing domains.
• Deep expertise in data modeling and ontology design, with experience unifying disparate data sources (e.g., weather, terrain, assets) into versioned, queryable systems.
• Strong background in MLOps and HPC/GPU workflows, including orchestration of batch and near-real-time pipelines on AWS, with a focus on reproducibility and cost governance.
• Comfortable collaborating closely with scientists and translating frontier research (e.g., ML diffusion downscaling, WRF-SFIRE) into production-grade, scalable systems.

Tips for this application

• Highlight specific projects where you designed a data ontology or unified data model that became the foundation for multiple applications—quantify the scale and impact.
• Emphasize hands-on experience with AWS services relevant to scientific compute (e.g., Batch, EKS, SageMaker, Step Functions) and cost optimization for GPU/HPC workloads.
• Demonstrate familiarity with weather or wildfire data formats and challenges (e.g., NWP models, ensembles, uncertainty quantification) in your resume or cover letter.
• Show evidence of 0→1 architecture: describe instances where you made foundational technical decisions that scaled over years, not just maintained existing systems.
• Include links to open-source contributions, technical blog posts, or talks that showcase your thinking on data platforms, MLOps, or scientific computing.

What to cover in your cover letter

['Your experience designing and owning a data ontology or unified data model that served as the backbone for complex, multi-source systems.', 'Concrete examples of architecting end-to-end platforms on AWS that handle large-scale scientific data, including orchestration of GPU/HPC workflows and cost management.', 'Your ability to collaborate with scientists and translate research into production systems, ideally in weather, climate, or geospatial domains.', "Why you are excited about OnPeak's mission and the challenge of building the technical foundation for weather decision intelligence in high-consequence industries."]

Research before applying

• Study OnPeak Intelligence's product and target customers (utilities) to understand the specific decision products they deliver and the data challenges involved.
• Research common weather and wildfire data sources (e.g., HRRR, GFS, WRF-SFIRE) and the technical challenges of downscaling and uncertainty quantification.
• Investigate the utility industry's asset data models (spans, poles, circuits, feeders) and how weather risk is currently managed to identify pain points OnPeak addresses.
• Look into recent advances in ML for weather downscaling (e.g., diffusion models) and deep learning weather prediction to speak credibly about frontier science.

Likely interview topics

Based on the job description, expect questions about:

• How would you design a data ontology to represent weather, fire, terrain, fuels, utility assets, forecasts, ensembles, uncertainty, and decisions in a coherent, versioned system?
• Describe your approach to orchestrating a pipeline that ingests continental NWP data, runs downscaling and fire-spread models, and delivers decision products to customers—what AWS services would you use and why?
• How do you ensure reproducibility and cost governance in GPU/HPC workflows for ML and scientific simulations?
• Walk us through a time you made a foundational architectural decision that had long-term implications. What was the context, and how did you validate it?
• How would you collaborate with the science organization to productionize a new ML downscaling technique while maintaining reliability and scalability?

Common mistakes to avoid

• Focusing only on past managerial or team-lead roles—this is an individual-contributor position that requires hands-on coding and architecture.
• Being vague about data modeling experience; avoid generic statements like 'designed databases' without specifics on ontology, versioning, and queryability.
• Ignoring the scientific domain—candidates who show no interest or curiosity about weather, fire behavior, or utility operations may be seen as a poor fit.
• Overemphasizing experience with standard web or mobile apps without connecting it to large-scale data platforms, MLOps, or scientific computing.