Software Engineer, Agentic AI

AboutKarbone Research is now Noreva. Noreva empowers energy and commodity producers, buyers, developers, traders, and investors to value assets, settle deals, stress-test capital structures, and build confident strategies - all in one customizable platform.We’re hiring an entry-level Software Engineer to join the team building Noreva’s full-stack forecasting platform, including new agentic AI tooling. This is a full-time role based in our Midtown Manhattan office.ResponsibilitiesHelp build and own components of the AI modeling agent that orchestrates the agentic workflow, including managing datasets, creating context, and building and updating modelsDesign and run experiments and proofs of concept, and help carry the strongest ideas into productionTest, evaluate, and improve agent performance, and build the tooling to measure it over timeContribute to the data pipelines and infrastructure that power the platformWrite clean, documented, testable code and take part in code review, both giving and receiving feedbackWhat You’ll LearnHow to design, evaluate, and scale agentic AI systems applied to real commodity price discoveryEnergy market fundamentals: power, capacity/resource adequacy, Renewable Energy Credits (RECs), and long-range forecastingHow analyst expertise becomes automated research and modeling workflowsHow to work closely with engineering, quant research, and product teams on a small team where your work ships quicklyQualificationsBS or MS in Computer Science, Engineering, or a related fieldSoftware engineering experience, including internships or substantial project workStrong Python skillsExperience with SQL and data tools, including PostgreSQL and TimescaleDB/TigerData for time seriesFamiliarity with data pipelines and cloud infrastructure (AWS or GCP)Experience with, or strong interest in, LLM and agent frameworks (function calling, RAG, LangChain or similar)Experience with vector databases is nice to haveComfortable taking ownership of problems and shipping production code on a small team