Conservation Metrics, Inc. (CMI) is a mission-driven social enterprise dedicated to providing technology-driven monitoring services that improve outcomes in the environmental and social justice spheres. The Lead Software Engineer, Data Sovereignty will own the architecture and data infrastructure behind the Guardian Connector platform, supporting Indigenous data sovereignty and community efforts through the development of innovative software solutions.
Responsibilities:
- Improve the stack-deploy developer experience Today, spinning up a new instance involves long playbooks and scripts: provisioning Auth0 tenants and app keys, installing apps on CapRover, and wiring environment variables across services (this is mostly not a problem Terraform can solve). You will own and dramatically simplify the installation and first-time configuration of the Guardian Connector software stack. Improvements here benefit both external users who run their own instances and our internal team, which hosts the stack on behalf of partners
- Design and implement governed data access systems Guardian Connector runs on a single VM with an on-box SQL database and media storage. Partners want interfaces that enable controlled data sharing across internal and external consumers, including with third-party BI, GIS, and AI tools. You’ll help to design role-based access controls (RBAC), implement scoped or time-bound data credentials, and define clear boundaries between private, shared, and public data. Supporting this pillar is the metadata and data catalog work needed to make data discoverable and understandable to the people granting and using that access
- Develop AI and LLM capabilities that preserve data sovereignty Rather than baking specific AI features into Guardian Connector — which would make our team a bottleneck on every new model and tool — we believe the more flexible path is to let partners bring the AI tool of their choice to their own data, with those tools acting as consumers of the governed access layer above. You’ll prototype and evaluate approaches such as local or self-hosted models, retrieval pipelines, and natural language interfaces that allow users to query and explore their data without external data leakage, and recommend one or two third-party tools we can endorse for partners
Requirements:
- You have configured PostgreSQL roles, grants, and RLS for production applications
- You've operated services in self-hosted or single-VM software stacks. For the most part this is not cloud-native work
- Your familiarity with Docker is at a level where you understand volumes and port mappings (the Guardian Connector stack uses CapRover, which runs apps on Docker Swarm)
- You've worked with identity systems (Auth0, Keycloak, or similar)
- You write design docs as a mode of working through architectural questions
- Empathy for non-technical users. You craft solutions for smart, fast learners who have limited experience with computers or smartphones. A listen-based, human-centered design process is at the heart of our work
- You can make and defend architectural calls under hard trade-offs, and you can do so collaboratively, bringing partners and teammates into the decision
- While LLMs are part of our toolkit (governed by human-first contributing guidelines), your work must be grounded in your own deep understanding of systems and sound technical judgement
- No specific formal educational credential is required. Relevant work history that demonstrates your growth and interest as a technologist is highly valued