AI Product Analyst

Salary:** 3000–4500$ gross 🏠 Format: remote work 🗣 English: C1

🎯 Responsibilities: – Conduct discovery sessions with the client's business owners, collect and agree on a catalog of questions for use cases with expected answers. Based on this, together with the client's analysts, form an immutable hold-out benchmark for go-live decisions. – Perform data discovery across sources (Oracle EPM, Landfolio, AcQuire): what is already in ADLS and Databricks, in what format, with what update frequency. Prepare a "reuse or build pipeline" decision and agree on it with the client's data team. – Describe the schema catalog and data quality checks. – Formulate requirements for the chat and dashboards: metrics, intents (retrieve, compare, rank, explain, verify), access rules, report templates (cost summary, license cost), document templates (pledge drafts, CMI report sections) and translate them into specifications for ML engineers. – Prepare a knowledge base for RAG: SOP, Tenement Management Guidebook, Mining Investment Law and Regulations; annotate datasets for evaluating retrieval and citation accuracy. – Conduct QA and benchmarking: run the harness, calibrate the LLM-as-a-judge rubric with the client's business lead, analyze errors (incorrect numbers, hallucinations, access violations), support UAT and the pilot group. – Maintain documentation: input into the architectural document, question catalog, knowledge transfer package, materials for weekly status and milestone demos.

🛠 Requirements: – At least 3 years of experience as an analyst (data / system / product analyst) on projects with data or ML for enterprise clients. – Proficient SQL: complex queries, reconciling numbers with dashboards and reports, data profiling. Experience with Databricks / Spark SQL or readiness to learn quickly. – Understanding of how LLM products work: RAG, Text2SQL, prompts, model limitations, hallucinations. Experience in forming test sets and evaluating the quality of responses (eval sets, rubrics, metrics). – Working English no lower than C1: workshops with client experts, correspondence and documents in English. – Independence in a distributed remote team, readiness for regular calls with the client in the Saudi Arabian time zone (UTC+3).

⭐️ Advantageous: – Azure ADLS, Databricks Genie, AI Foundry, Power BI, or similar BI tools. – Experience with financial data: budget, actual, forecast, P&L. Familiarity with Oracle EPM / Hyperion. – Experience with geological data, AcQuire, GIS (ArcGIS and similar), spatial data. – Domain: mining, oil and gas, heavy industry. – Python / pandas for data profiling, experience with LLM evaluation tools (LLM-as-a-judge, Ragas, promptfoo, and similar).