I'm a Product Engineer with a background in Geospatial Information Engineering and Civil Engineering, currently working at a satellite-manufacturing company. I enjoy owning the full lifecycle of a product — from screen planning, use-case extraction, and DDD-based backend modeling, to implementation, E2E testing, and deployment.

My happiest engineering days are when I can compress what would normally take a 6-person team (PM · designer · frontend · backend · infra · QA) into a single coherent flow — without dropping quality. That's the lens I bring to every project.

• 🛰️ Domain expertise in satellite imagery, GIS, geodesy, and SAR signal processing (Sentinel / Umbra / Kompsat, DInSAR / PSInSAR / SBAS via SNAP)
• 🧩 Comfortable across the stack: React/Next.js on the frontend, NestJS/PostgreSQL on the backend, Terraform/AWS for infra
• 🧪 Strong believer in full-cycle testing — Playwright E2E + IaC-driven integration tests as a project safety net
• 📱 Ship PWA-based multi-platform apps (desktop + mobile) from a single codebase
• 🤖 Currently exploring AI/LLM integration to make products more usable and accessible

Backend Developer & Product Engineer · Nov 2025 – Present Frontend Developer · Nov 2023 – Nov 2025

End-to-end ownership across multiple satellite-domain products: screen planning → use case extraction → DDD backend design → implementation → E2E testing → deployment.

🛰️ Satellite Analysis Platform A platform that ingests imagery and metadata from multiple satellite providers (Sentinel, Umbra, Kompsat, etc.), normalizes them into a queryable DB, and lets users draw an AOI (Area of Interest) to search, request, and review analyzed products. Bridges my Master's research in geospatial information directly into a production product — including 2D satellite imagery visualization tied to Sentinel pipelines.

📡 SDPE — SAR Data Processing Element Backend pipeline service that takes raw SAR signals from Lumir's own SAR satellites and produces final products. Built the orchestration backend plus a N8N-style pipeline editor on the frontend so operators can compose and manage processing flows visually. Owned spec, design, and implementation across both layers.

🛰️ Sentinel-1 InSAR DSM Extraction Pipeline (in progress, 2026) Government-collaborated R&D project building an automated InSAR pipeline that extracts a DSM (Digital Surface Model) from Sentinel-1 SLC pairs as a temporary reference DEM for a DInSAR surface-displacement monitoring task — used until the national 5m DEM becomes available.

• Stack: ESA SNAP 12 GPT (-Xmx87G) + SNAPHU 2.0.3 orchestrated from Python (subprocess) on Ubuntu 20.04 (64 CPU / 125 GiB RAM / NVMe), with GeoPandas/Shapely/PyProj/Rasterio for AOI + raster handling and QGIS for visual QA
• Pipeline (Stages 0–11): TOPS-Split → Apply-Orbit-File → Back-Geocoding → Enhanced Spectral Diversity → Interferogram → Deburst → Goldstein filter + Multilook → SNAPHU Export/Unwrap/Import → Phase-to-Elevation → Range-Doppler Terrain Correction → GeoTIFF (WGS84/UTM 52N)
• Engineering choices: every stage is independently re-runnable from prior artifacts; all parameters (AOI, baseline thresholds, multilook factors, master/slave product IDs) externalized to pipeline_config.yaml so the pipeline generalizes to other AOIs; intermediates kept on local NVMe (SMB write-avoidance) to dodge I/O bottlenecks
• Migration gotcha solved: SNAP 12 renamed S1TBX → MicrowaveTBX (eu.esa.microwavetbx.*), so InSAR operators (TOPSAR-Split, Back-Geocoding, SnaphuExport, …) aren't enabled by default — documented the --modules --install recipe to reproduce a working CLI environment
• Hands-on with the broader InSAR toolkit — DInSAR, PSInSAR, SBAS — being folded into the Satellite Analysis Platform as the analytics backbone

📚 Internal Backoffice / CMS Built a content management system covering company announcements, training materials, and an internal wiki — backend service ownership.

Research focus: Geodesy · Surveying · Satellite Imagery · GIS

A multi-platform (desktop + mobile PWA) household inventory app — track expiration dates, get smart restocking suggestions, and share inventory across multiple households with role-based permissions. Designed, built, and operated solo — replacing what would normally take a full PM/Design/FE/BE/Infra/QA team.

Highlights

• 🏘️ Multi-household & role-based sharing — invite users with granular permissions per household (home, office, vehicle, …)
• 📦 Hierarchical space model — Room → Furniture → Storage Slot for precise location tracking
• ⏰ Expiration & lot tracking — per-lot expiry, automatic warning notifications based on remaining shelf-life
• 🛒 Smart shopping suggestions — combines low-stock + near-expiry signals
• 🤖 AI roadmap (in progress) — LLM chatbot for product usability, voice input, and Tesseract-based image recognition
• ⚙️ Full-cycle delivery — DDD-driven use case design dramatically simplified the dev process and saved time

Stack

Engineering practices I'm proud of here

• ✅ Frontend E2E coverage with Playwright as a stability net for solo development
• ✅ Terraform-driven IaC integrated into both backend and E2E tests — full-cycle testing where infra, API, and UI are all verified together
• ✅ PWA from day one — same codebase ships to desktop and mobile
• ✅ Use-case-first development — fewer meetings, fewer hand-offs, fewer bugs

🚧 Sibling project (in planning): A shared household ledger app — collaborative budgeting with the same multi-user/role-based foundation.