I build AI products that turn messy enterprise, GTM, and operational signals into trusted workflows, scored decisions, and measurable actions. My work sits where product strategy meets applied LLM systems: defining the decision the product is making, grounding it in real data, and shipping the evaluation and tooling that make the output trustworthy.
- Agentic workflows that take raw, unstructured signals and produce scored, explainable decisions
- Evaluation and scoring systems that make LLM output measurable and reviewable
- Build-vs-buy and architecture tradeoff analysis across cost, latency, reliability, and operating complexity
- Automation that removes manual steps from knowledge and operational workflows
An AI product that turns energy and operational signals into clear, actionable guidance for users. Built and shipped as a live product.
- Live: trywattwise.com
- Repo: wattwise-ai
A platform that ingests fragmented investor and LP signals and converts them into structured intelligence for relationship and pipeline decisions. The full platform is private; a demo is available publicly.
- Demo repo: lp-intelligence-demo
An LLM-as-judge evaluation tool for scoring model responses. EVALCARD scores output across five dimensions — factual accuracy, completeness, clarity, depth, and relevance — against a 50-point rubric, and exports results as JSON or Markdown for review and tracking.
- Repo: evalcard
A structured framework for deciding whether to integrate existing LLM APIs or build custom models for a given AI feature. It frames the decision as a tradeoff analysis across cost, latency, reliability, and operating complexity, so product teams can reason about architecture choices with clear criteria.
- Repo: ai-buildorbuy-framework
A tool that automates discovery and prioritization of relevant roles. It analyzes job descriptions, compares them against a candidate's background and goals, runs on a schedule to capture new postings, and ranks results by fit to cut manual search effort.
- Repo: ai-job-discovery-engine
An automation tool that exports course materials from a learning management system and organizes them into structured Google Drive folders, replacing hours of manual downloading and filing with a repeatable pipeline.
- Agentic systems and applied LLM workflows
- LLM evaluation and scoring
- AI product strategy and architecture tradeoffs
- Automation and operational tooling
- WattWise: trywattwise.com
- LinkedIn: teja-padala
- GitHub: tejaswarpadala-a11y