Founding Software Engineer – VP of AI Voice

The Business (B2B2C & The Factory of Verticals): macafe.cloud is not just another SaaS; it is the new operating system for physical security. We operate a powerful B2B2C machine. By empowering businesses with zero-trust video and forensic acoustic infrastructure, we enable them to sell absolute transparency and peace of mind to their consumers. This is a factory of verticals, and your job is to dominate them one by one.

The Reality Check: Let’s cut the noise. I am a Hacker. I spent the last few months in the trenches, bootstrapping the cloud-native middleware that makes this vision possible. The architecture is built, the servers are running, and the code is lethal. The technical risk is solved. Now, I am ready to share the cap table with a Go-To-Market shark who can turn this infrastructure into pure revenue.

The Mission: Video is directional and heavy. Audio is omnidirectional, lightweight, and detects intent before the action happens. You will own our entire Audio Intelligence Layer. Your mandate is to build a highly resilient microservice architecture that ingests continuous audio streams (e.g., from .dav or .aac containers), extracts intelligence, and structures the data for our immutable governance engine. You aren't just scripting ffmpeg; you are building a fault-tolerant pipeline capable of handling 10,000 concurrent streams without dropping a frame.

Your Core Mandate:

• Signal Extraction (The Physics): Build the engine that demuxes audio, normalizes it (PCM, 16kHz), and calculates volume telemetry (RMS/dBFS) over sliding windows to generate a second-by-second time series.
• Transient Detection (The "Hit"): Implement low-latency transient onset detection algorithms. You must differentiate between ambient noise and sudden structural impacts (Hits/Strong Hits) by analyzing amplitude deltas (< 50ms attack time) and spectral energy (FFT) in the sub-150Hz range.
• Semantic Intelligence (The Context): Integrate and optimize Speech-to-Text (STT) models (e.g., local Whisper deployments) to transcribe speech, detect anomalous noise (screams, alarms), and extract intent.
• Resilience at Scale: Apply "Fintech-level" rigor to physical security. Handle corrupted audio frames gracefully and structure the codebase for extreme observability and failure resistance.

The Culture (Zero Ego Policy):

• You are a Leader, not a "Boss": We do not tolerate blame games, corporate politics, or ego boundaries. You solve problems, you don't point fingers.
• Extreme Execution: You are a maker. You aren't afraid to dive deep into codec variations (AAC, G.711A), optimize local AI models, and debug raw waveforms.

The Deal (High Risk, Massive Reward): This is a bootstrapped, Day 1 startup environment.

• The Offer: Sweat Equity (Partnership).
• The Vesting: A strict 4-year, back-loaded vesting schedule. The "gold" is at the end of the journey to ensure absolute long-term commitment.
• The Cliff: Year 1 is a brutal, performance-based cliff. If we do not hit our architectural and throughput milestones by month 12, the contract terminates, and you walk away with 0%.
• Want more equity? Put cash on the table. If you want a traditional 15% or 20% Co-Founder slice, you must buy in at our current pre-seed valuation. No free lunches.

Who You Are:

• Fluent in English (Mandatory).
• You have a deep background in high-performance computing (Python, Java, Go, or C++) and are comfortable with audio signal processing (FFT, RMS, Transients).
• You build systems that don't fail when the network stutters.