Geetesh Veeravalli

GPA 3.6 / 4.0 · Dean's List

Coursework Data Structures & Algorithms, Operating Systems, Software Engineering, Programming Languages, Computer Organization, Information Assurance

1. AI Engineer I, Floodlight, Inc.

   Austin, TX · May 2026 — Present

   • Own the Climate Value-at-Risk (CVaR) engine behind 300+ client reports (annual loss, 95/99% VaR, max loss)
   • Shipped LLM-generated risk recommendations and per-hazard risk attribution across the dashboard and CVaR reports
   • Cut report generation from ~120 s to under 10 s with a report-caching layer and PostgreSQL query optimizations
   • Built a Model Context Protocol (MCP) server system exposing X-STILT atmospheric-model outputs to LLM agents
   • Integrated Vulcan, EDGAR, and ODIAC emissions inventories for Scope 1/2 reporting; built a Green Building Score
   • Launched credit-based pricing and payments and a rebuilt Next.js sign-up/onboarding flow, lifting completion ~30%
   • Led a technical audit of the risk engine and designed the GWL-indexed scenario-planner architecture
   • Shipped a default-deny Public Access Layer that serves unauthenticated demo reports to prospects
   • Deployed the new Next.js marketing site to a GCP VM behind Cloudflare (SSL/origin certificates, env config)
   • Resolved a production outage (runaway logs filled the disk, killing MySQL) and averted a destructive DB restore
   • Prototyped a global air-quality downscaling model (LightGBM, CMIP6 projection); co-authoring a Climate VaR paper

2. Software Engineering Intern, Floodlight, Inc.

   Austin, TX · May 2025 — May 2026

   • Automated six climate/energy ETL pipelines (n8n, GCS, PostgreSQL) with 100% on-schedule runs and alerting
   • Ported the open-source X-STILT transport model from R to typed Python 3.12, cutting simulation time ~35%
   • Added Smart Drill-Down and One-Click Suggest to the geospatial risk dashboard, cutting analyst time-to-insight 40%
   • Wrote the platform Data Dictionary v0.1 and the full data-source and methodology documentation

3. Data Scientist, ASU LightWorks

   Tempe, AZ · Nov 2024 — Present

   • Data modeling and digital twin lead on an 18-month, $2M USFS pilot proposal for a 1M+ acre biomass system
   • Built an interactive Leaflet.js digital twin of the biomass system: suitability overlays, coded nodes, corridor distances
   • Built the grant budget workbook and delivered-cost model pricing the biomass supply system for USFS funders

• DigiClips subscriber options, ASU Capstone · Software Lead

  Aug 2025 — May 2026

  i18n · UI overhaul · Live · digiclipsproject.com ↗

• IBSO biomass optimization model, ASU · Dr. Rimjhim Aggarwal

  Aug — Dec 2025

  Python · Optimization

• Climate Value-at-Risk engine, Floodlight

  2025 — 26

  Python · LightGBM · PostgreSQL · Production

• X-STILT MCP server system, Floodlight

  2026

  Python 3.12 · MCP · Anthropic + OpenAI APIs · Production

Languages

Python, TypeScript, JavaScript, SQL, R, C++, Java

ML

pandas, NumPy, scikit-learn, LightGBM, PyTorch

Climate + Geo

CMIP6, ERA5, CAMS, X-STILT, Leaflet.js

Platform

GCP (GCS, BigQuery, Compute), Docker, Cloudflare, PostgreSQL, n8n, Linux

Agents

Anthropic API, OpenAI API, Model Context Protocol, Claude Code