Applied AI Engineer
Our Technology
Terawatt is not a traditional infrastructure operator. Our technology platform transforms a portfolio of sites into a reliable, intelligent network — through advanced power management, operational optimization, and data-driven systems that back our customers with hard performance guarantees. That's what makes us a de-risked operator and a category-defining company.
We sit at a rare intersection: software, hardware, real estate, energy systems, and live site operations. The work is concrete and the stakes are real. The code you write today runs physical infrastructure tomorrow. You'll build systems that orchestrate megawatts of power across hundreds of chargers, ingest millions of telemetry events per day, and keep autonomous fleets, rideshare vehicles, and heavy-duty trucks moving — in real time.
Role Description
In this role you build the engine that turbocharges Terawatt’s ability to build, operate, and scale the infrastructure of the future. This is a very important role with direct support from Terawatt’s executive team.
As an Applied AI engineer, you will be the software engineer building Terawatt’s internal AI platform. You will partner with stakeholders to translate high-value, ambiguous business problems across development, finance, and operations into well-framed agentic problems with clear success criteria and evaluation methodologies. You will provide technical leadership across the full development and evaluation lifecycle, including post-deployment iteration, for agentic workflows and custom internal applications that leverage the latest innovations with our internal data sources to provide a competitive edge to our entire organization.
If you care about sustainability and you’re excited to use AI to accelerate clean energy infrastructure, this is your chance to be part of a rapidly-growing, well-funded climate tech scale-up.
Core Responsibilities
Architect the internal AI platform. Create the reusable primitives — evaluation harnesses, MCP servers, tool libraries, prompt patterns — that turn one-off prototypes into a flywheel
Embed with teams across the business. Work closely with Development, Finance, and Ops teams. Together with the PM for AI, translate ambiguous, high-stakes problems — site selection decisions, financial & layout modeling, interconnection workflows — into well-scoped agent specs with clear success criteria
Prototype and build agentic workflows. Design and ship LLM-powered agents that reason, plan, and act across Terawatt's tools and data. Start with prototypes, then iterate based on user feedback to production-grade reliability, security, and cost
Own Quality: Define and apply evaluation and quality standards to measure success, failures, and regressions. Debug real-world agent behavior and systematically improve prompts, workflows, tools, and guardrails making them a seamless and trusted part of the daily workflow for every Terawatt employee
Operate what you ship — instrument agents with monitoring and cost/latency budgets, catch quality regressions before users do, and keep workflows running as underlying models change
Rigorous Evaluation: Establish quantitative evaluation frameworks to assess agentic precision, operational safety, and latency performance — moving beyond heuristic testing to measurable outcomes
Deep Curiosity: You stay current with the fast-moving AI landscape, run experiments, and actively share findings and best practices with the team
You communicate clearly across technical and non-technical audiences and thrive in a collaborative, fast-moving hybrid environment
Minimum Qualifications
Production Engineering: 5+ years in architecting, deploying, and overseeing production-grade software using Python or TypeScript with a strong emphasis on writing code that is modular, observable, and built to scale
Agentic Architectures: Hands-on experience of engineering autonomous agents capable of multi-step reasoning and execution—utilizing patterns like ReAct or Plan-and-Execute—while integrating seamlessly with external toolsets, APIs and MCPs
Experience building production applications using LLM APIs (e.g., Claude, OpenAI) to enhance workflows or productivity across cross-functional teams.
The LLM Stack: Deep technical proficiency with frontier models, RAG implementations, and vector databases. Experience with orchestration frameworks such as LangGraph or CrewAI
Security First: Experience building AI platforms that treat security as a core feature, including familiarity with risks specific to agentic and LLM-based systems
Proficiency with Git workflows and agile development practices
Preferred Qualifications
Proven ability to work closely with Product to scope, prioritize, and independently deliver features from definition through deployment
Experience mentoring junior engineers or contributing to architectural design discussions
Experience with Git workflows, agile methodology, and contributing to architectural discussions
Ability to customize and retrain models for advanced internal use cases via practical experience in machine learning frameworks such as TensorFlow, PyTorch, or scikit-learn