AI engineer

Ekthar works at the intersection of wildlife protection and conservation. We are building an internal AI capability from scratch — a small, hands‑on team that will design and ship agents, automations, and internal platforms that make the rest of the organization faster, sharper, and more data‑driven. This is a builder's seat, not a research seat.

The role:

You will join a small founding team responsible for delivering AI‑powered solutions across Ekthar. You will own problems end‑to‑end — scoping, building, deploying, and maintaining the agents and platforms that support our mission.

• Build AI agents and workflows that solve real, named problems for our operations, field, research, and leadership teams.
• Ship internal platforms — small tools, dashboards, and agentic systems — end to end. You design it, you build it, you maintain it.
• Work fluently across n8n, Open Claw, custom Python/Type Script, and orchestration frameworks (Lang Graph, Claude Agent SDK, or similar).
• Decide pragmatically between cloud LLMs (Claude, GPT, Gemini) and local/self‑hosted models (Llama, Qwen, Mistral via Ollama, vLLM, or LM Studio) based on cost, privacy, and latency.
• Integrate with the tools the organization already uses (Google Workspace, Microsoft 365, GIS and conservation data sources, etc.).
• Help define the team's stack, standards, and roadmap. We expect strong opinions, loosely held.

• A builder. You'd rather ship a scrappy v1 in a week than write a 30‑page proposal.
• Technically deep — you can code (Python and/or Type Script at minimum), read APIs, debug a broken n8n node, and reason about model behavior.
• You’ve actually built something with LLMs that other people use — agents, RAG systems, automations, copilots. Side projects count. Strongly.
• Comfortable being early. No ML platform team, no dedicated Dev Ops, no playbook. You build the playbook.
• Curious about wildlife, conservation, ecology, or climate — or curious enough to learn fast.
• 2+ years writing production code (Python or Type Script/JavaScript).
• Hands‑on experience with at least one LLM API (Anthropic, OpenAI, Google, or open‑weight via Ollama/vLLM).
• Experience building agentic or multi‑step workflows — not just one‑shot prompts.
• Familiarity with n8n, Make, Zapier, or comparable automation platforms (n8n strongly preferred).
• Comfort with Git, REST APIs, and basic cloud infrastructure (AWS, GCP, Azure, or self‑hosted.