Geospatial Data Scientist working across LiDAR, aerial imagery, computer vision, and agentic AI systems — from field survey through production platforms supporting critical utility infrastructure.
I work at the intersection of geospatial engineering, applied AI, and cloud data platform engineering — currently as a Data Scientist at The Spear Group, technical lead across Florida Power & Light's AI-based infrastructure inspection program. I set technical requirements, build the data pipelines and cloud infrastructure behind the program, and serve as subject-matter expert through vendor model development. I also build the GIS applications used for infrastructure monitoring and disaster response — including a portal used during Hurricane Milton.
My background started in the field: GNSS and UAV survey, municipal geodatabase design, and control surveys as a Geomatics Engineer in Nepal — before graduate research at Florida Atlantic University on flood hazard modeling and environmental justice sharpened the AI/ML and geospatial data engineering side. That research led to a peer-reviewed publication on flood hazard exposure and social vulnerability in the U.S., and I'm a co-author on PyWMP, an open-source watershed modeling framework, with the accompanying manuscript currently under peer review.
Outside of work I contribute to open-source geospatial and AI tools — a published Python package for road-centerline extraction, agentic AI systems for querying spatial and hazard data, and a flood/storm-surge exposure pipeline for utility assets, among others below. I'm FAA Part 107 certified.
Technical lead across FPL's AI and geospatial infrastructure inspection program — from vendor selection through data engineering to production deployment.
Open-source contributions built outside of my employment — agentic AI over spatial/hazard data, a published Python package, and full data-to-map pipelines. Most ship with a live demo; nothing here requires an API key to try.
A published Python package that extracts road centerlines from polygon geometries — worldwide, in any input CRS, in any format GeoPandas can read. CLI and Python API, computed via medial-axis skeletonization.
An MCP server unifying FEMA, USGS, NOAA, NWS, and NASA hazard data behind a single AI-agent interface — one-call location risk summaries and map rendering, compatible with Claude Desktop.
A cloud-native pipeline overlaying NOAA storm-surge and flood-inundation data against Overture Maps building footprints, scoring per-asset flood exposure across 8 coastal regions — with a map and API.
Generates DEM, DSM, hillshade, slope, aspect, and contour products from USGS 3DEP lidar. Draw an area of interest in the browser; it generates a configured local script — no data ever uploads.
Ask a plain-English question about San Francisco's utility infrastructure — a local LLM agent (LangGraph ReAct, Ollama) translates it to spatial SQL, runs it, and answers with a summary and interactive map. No API keys.
Converts a spreadsheet of US street addresses into geocoded points and building footprint polygons using free Nominatim and Overpass APIs — as a no-install web app or a scriptable Jupyter notebook.
A LangChain RAG chatbot with a history-aware retriever for multi-turn, context-grounded conversation — every answer grounded in retrieved source material, with citations. Runs entirely locally via Ollama.
A generalized data-engineering and analytics platform for GTFS transit feeds — ingest any agency's schedule data, measure headways, stop coverage, and spacing anomalies, and explore results in an interactive map app.
Converts romanized Nepali — including inconsistent, texting-style spelling — to Devanagari live, entirely client-side. A CLI batch-translates SRT subtitles through pluggable backends with caption-merging normalization.
Production AI needs infrastructure, data pipelines, and internal tooling around it. This is a selection of the platform-layer components I've built to support enterprise geospatial and AI systems.
AWS CDK component framework with enterprise naming, tagging, and deployment abstractions for ML and app workloads.
Modular platform for taxonomy alignment, schema publishing, quality validation, and operational patch workflows (Python, SQL, DuckDB).
CLI and web console for environment-aware API operations with shared authentication and endpoint tooling (FastAPI, React, Typer).
PostgreSQL schema migration CLI with source/target connection separation, dry runs, and preflight checks.
ONNX-based object detection package for utility equipment with COCO JSON output and optional visualization.
Reusable package for API extraction, taxonomy joins, metadata patching, and downstream publishing automation.
Boundary-focused architecture separating API adapters, data sync logic, and publishing interfaces from generic scripts.
Modular script layer for analysis, checks, loaders, fixes, and patch orchestration with migration-safe wrappers.
Optional FastAPI sidecar for heavier geospatial processing, conversion runtimes, and algorithm endpoints.
Focused API exposing health, pagination, and entity summary endpoints for analytics demos and stakeholder review.
Joins 8M+ drone inspection images with 350K+ historical condition assessment reports via spatial and temporal joins, writing enriched metadata back through batched API calls.
Streamlit app letting internal teams and AI vendors pull training images by asset, condition, or class across 25+ asset types and 400+ condition classes — instead of searching the full image set by hand.
Automated integration of image metadata and model-generated detection annotations into an Esri PostGIS geodatabase, so vendor model output can be analyzed geospatially.
ArcGIS Experience Builder/Dashboards portal used during Hurricane Milton to track storm impact, restoration progress, and crew deployment in real time.
Data science workspace scaffold for experimentation, feature engineering, and model lifecycle workflows.
Structured runbooks, standards, acceptance gates, and deployment checklists for end-to-end delivery governance.
A small browser-side terrain model in the spirit of the FEMA-based hydrologic exposure work below — procedurally generated elevation, not the published dataset. Drag the slider to raise sea level and watch inundation spread.
Color ramp: low-lying green → upland tan/white. Blue = below current sea level.
Utility corridor: 0 of 7 poles below sea level
Most of my professional work is confidential to the organizations I've supported — publications below are what I can share in full.
MSc, Geosciences — GIS Concentration · Florida Atlantic University · 2023–2024 · GPA 4.0/4.0
BSc, Geomatics Engineering · Tribhuvan University, Nepal · 2016–2021 · GPA 3.64/4.0
Open to Data Science, GIS/Geospatial Engineering, and Data/Platform roles. Based in West Palm Beach, FL.