Tech Stack
Job Description, Responsibilities & Requirements
About the Position
Native / Desktop Engineer
The role. Build Desktop Intercept - the native Windows driver and app that acts as a virtual mic/speaker, replaces audio in real time, and (Phase 1 Phase C) reads the DOM and drives Computer Use. This is the hardest engineering on the platform and the deepest moat. Everything you build feeds a high-efficiency, event-driven, low-latency distributed platform - voice, desktop, intelligence, and AI combined - so you think in latency budgets, throughput, and efficiency, not just features. Startup environment: 1-week sprints, Friday demos, fail fast and move forward.
What you'll own. Windows native driver (WASAPI, virtual audio devices, kernel-adjacent hooks) · real-time bidirectional audio routing · the user-space app (audio router, DOM scraper, MCP server) · the desktop-to-cloud path onto the event bus (NATS-class, WebSocket) under a sub-300ms budget · AV co-existence and mass-deployment · your committed timelines.
Who you are. Self-starter, hacker, grit. You consider yourself exceptional at the systems level and can show it. You love new technology, adapt fast when it changes, demonstrate ideas with working code, and use AI tools daily to multiply your velocity. Team player who makes and hits committed timelines. Loves solving the problems nobody else can - and winning.
Responsibilities
- Build a native Windows driver and app for real-time audio routing and DOM scraping.
- Develop real-time bidirectional audio routing.
- Create the user-space app (audio router, DOM scraper, MCP server).
- Establish the desktop-to-cloud path onto the event bus (NATS-class, WebSocket) under a sub-300ms budget.
- Ensure AV co-existence and mass-deployment.
- Meet committed timelines.
Requirements
- Mandatory:
- 8+ years systems / native development; shipped a Windows desktop agent at real scale (>10k seats) with audio.
- Strong C/C++ skills.
- Strong Go skills for the service side.
- Comfortable with:
- Building and interacting with drivers - virtual audio devices, virtual microphones/speakers, WASAPI / Core Audio, kernel-adjacent hooks.
- Real-time audio pipelines under a sub-300ms budget - buffering, resampling, drift, glitch-free capture/playback.
- High-efficiency systems end-to-end - event-driven architecture, pub/sub, WebSocket streaming, and understanding state, race conditions, and locking at the native layer (lock-free audio buffers, real-time threads, priority inversion).
- A plus if you have:
- macOS / Linux native experience.
- Browser / DOM internals and process injection experience (Phase C).
- Enterprise mass-deployment craft - MSI/installer engineering, silent deploys, fleet upgrades across 10k+ locked-down seats.
- Driver/code signing and AV/EDR co-existence - WHQL, EV certs, and staying invisible to CrowdStrike-class agents without tripping them.
- Crash-dump and telemetry discipline - WinDbg/minidumps, symbolized crash reporting from the field; you debug machines you'll never touch.
- Security posture - least-privilege service design; this app hears customer audio and reads screens, so it must be provably tight.
- A master debugger at the systems level - kernel logs, crash dumps, audio glitches, AV interference; you see it.
- SIP / RTP experience.
- Mid-tier or better hands-on AI-in-product experience (the desktop app hosts an MCP server and drives Computer Use).
We Offer
- Opportunity to work on cutting-edge technology.
- Collaborative startup environment with 1-week sprints and Friday demos.
- Fail fast and move forward culture.
- Competitive compensation and benefits.
About the Company
Alcor is at the forefront of developing high-efficiency, event-driven, low-latency distributed platforms that combine voice, desktop, intelligence, and AI. We are committed to pushing the boundaries of technology and creating solutions that solve complex problems.