HW

Martin Mikušát

AKA Hal Wayland

Full-Stack Developer at Delta Electronics (via W&K Industrietechnik s.r.o.)

Location: Trenčín, Slovakia Email: m.mikusat@gmail.com Portfolio: halwayland.com GitHub: github.com/MartinMikusat LinkedIn: linkedin.com/in/martin-mikusat Languages: English (C2), Slovak (C2)
Professional Summary

Full-stack developer who independently designed, built, and owns a production EV charger diagnostics platform used daily by technicians across EMEA, supporting ~5,000 chargers. Focused on performance, UX, and shipping complex systems end-to-end. Experienced with Electron, Svelte, Supabase, large-scale log processing, and LLM-assisted development in production.

Professional Experience

Full-Stack Developer

Delta Electronics — Global power electronics and EV charging infrastructure company

Via W&K Industrietechnik s.r.o. — Dubnica nad Váhom, Slovakia | March 2022 – Present

  • • Initiated and architected an Electron/Svelte desktop app for EV charger diagnostics, used daily by technicians across EMEA supporting ~5,000 cabinet chargers
  • • Built web worker-based log processing pipeline: decompresses, parses, and deduplicates Charge.log files into structured charging sessions with ~150ms parsing and instant filtering, keeping UI responsive during CPU-intensive operations
  • • Implemented multi-type search (chargers, cases, work orders, spare parts, knowledge articles) with fuzzy matching, exact serial prioritization, and 21+ typed filters, delivering sub-200ms results across tens of thousands of records
  • • Designed offline-first architecture with Dexie local caching, real-time Supabase synchronization, and readiness gates that coordinate data loading, heuristic online checks, and realtime subscriptions before app entry
  • • Built automated L2 handoff workflow: discovers critical logs, downloads via queued operations, gathers from cache with freshness reuse, and packages diagnostic data into zip files — reduced preparation time from ~5 minutes to ~30 seconds per case, saving approximately 2 hours daily across ~30 daily escalations
  • • Implemented LLM-powered features: case/email summarization to reduce manual workload, and fuzzy location matching for work order validation using web worker-based processing with incremental result updates
  • • Delivered fleet-wide analytics and reporting with streaming XLSX exports designed to scale to thousands of chargers — enabling fleet-wide visibility that would not exist without the application
  • • Built bulk action system for batch operations (connection checks, log downloads) across hundreds of chargers simultaneously, replacing manual per-charger workflows that would be impractical at scale
  • • Implemented real-time connection health monitoring and power-user workflows (paste-to-search, click-to-copy, instant feedback)
Personal Projects

Parametric Animation Sequencer (TypeScript)

Framework-agnostic TypeScript library for defining parametric 3D animation sequences — like CAD but for animation. Supports relative timing dependencies.

  • • Designed and implemented a functional, reconciliation-based API that converts declarative animation graphs into resolved keyframes and deterministic animation state
  • • Implemented time reconciliation for complex relative timing dependencies (absolute times and offsets relative to parent start/end)
  • • Built tree-shakeable 3D API to optimize bundle size and maintain clear abstraction boundaries
  • • Implemented mathematically correct interpolation strategies: linear interpolation for positions, quaternion SLERP for 3D rotations
  • • Designed camera animation system for 3D (rotation, target, zoom) scenes
  • • Implemented marker-based hierarchical transform positioning for 3D scenes
  • • Created a documentation pipeline that generates API documentation from TSDoc comments using TypeDoc and a static documentation site workflow

Technologies: TypeScript, Animation Systems, Linear Algebra, Quaternions, Tree-Shaking, TSDoc, TypeDoc

Odin Standard Library Search (SvelteKit)

Work in Progress

Real-time search interface for the Odin programming language's standard library documentation. Provides fast access to base, core, and vendor packages with offline-first architecture.

  • • Multi-stage search pipeline: exact match, metadata match, and fuzzy search (Fuse.js) via Web Workers with 150ms debouncing
  • • Offline-first with IndexedDB caching (Dexie) and incremental sync from Supabase PostgreSQL
  • • Data pipeline: scrape script generates Markdown, generation script processes into Supabase with deterministic UUID-based upserts

Technologies: SvelteKit, TypeScript, Supabase, Dexie, Fuse.js, Web Workers

Clockwork 2 Interactive Assembly Manual (Three.js)

Work in Progress

Interactive 3D step-by-step assembly manual for the Voron Clockwork 2 extruder. Animated 3D models guide the user through each assembly stage with playback controls and chapter navigation.

  • • Modeled, UV-unwrapped, and exported all 3D assets using Fusion 360, Blender, and RizomUV — full pipeline from CAD source to web-ready GLB
  • • Built with Three.js via Threlte (Svelte bindings), using the Parametric Sequencer library for declarative animation definition
  • • Features animated camera, chapter navigation, parts list sidebar, and playback controls (play/pause, scrub, speed)

Technologies: SvelteKit, Three.js, Threlte, TypeScript, Fusion 360, Blender, RizomUV, parametric-sequencer

Key Technical Skills
Programming: TypeScript, JavaScript, Python (familiar)
Frontend/Desktop: Svelte, React, Electron, HTML, CSS, Tailwind CSS
3D/Animation: Three.js, Threlte, custom animation systems
Backend/Data: Supabase (PostgreSQL, Auth, Realtime), Dexie (IndexedDB), log parsing/data processing
Testing/Tooling: Vitest, Git, Vite, Bun
Performance/UX: Low-latency search/filtering, large dataset rendering, Web Workers, power-user workflows
AI/LLMs: LLM-assisted development (Cursor, Codex), production summarization/automation
Exploring: Odin, ReScript, Rust
Education

Gymnázium Ľudovíta Štúra - Trenčín, Slovakia

Maturita - Informatics and Mathematics

Certifications

Deep Learning Specialization - Coursera (Andrew Ng)