Stephen Pearse — CV (March 2026)

pearse.audio@gmail.com · LinkedIn · GitHub · spearse.github.io/portfolio

Profile

Audio/AI engineer and former academic with 13+ years building tools for music creators. Led cross-platform DAWs, AI arrangement systems, semantic audio search, and payments platforms. Comfortable owning architecture, delivery, and team patterns from research prototypes through production releases.

Core Skills

Audio/DSP: C++/JUCE, Qt/PyQt, <10ms real-time pipelines, LUFS/EBU R128, filters/effects, hardware integration (CDJ/ASIO/CoreAudio), cross-platform delivery.
AI/ML systems: Multi-agent AI (LangGraph), voice AI (ElevenLabs Conversational AI), LLM integration (Vertex/Gemini), cloud AI deployment (LangGraph Platform, Google Discovery Engine), semantic search (MERT, CLaMP3, FAISS), streaming parsers/DSLs for musical structure, PyTorch/Transformers.
Full-stack/infra: TypeScript (NestJS, React, TypeORM), Python/FastAPI, Chargebee/webhooks, CI/CD (GitHub Actions, signing/notarization, Windows installers), Docker/GCP, feature flags/observability (ConfigCat, Sentry).
Leadership: Tech lead, interviewer, onboarding/patterns, cross-functional collaboration with ML/backend/product, rapid delivery under tight timelines.

Experience

Density (Audio/AI tools) — Lead Software Engineer / Tech Lead / AI Design Lead (2021–Present)

Cross-platform DAW: Led Mac-only → Mac/Windows port; reworked threading/interop; built Windows CI/CD; universal binaries/signing; crash rate down 40% (5%→3%); release cadence from monthly to ~1.5 weeks; supported ~2,500 DAU.
Real-time audio/DSP: Lock-free messaging, preallocated buffers, LUFS metering (BS.1770-4), filter modeling, fractional delay/multi-band limiting, hardware recording from CDJs.
AI Copilot (voice assistant): Sole owner of Electron + React 19 + TypeScript voice interface (ElevenLabs speech-to-speech, Three.js 3D orb, custom interrupt system); heavy contributor to LangGraph multi-agent backend (intent routing, palette curation via Google Discovery Engine, DensityMark arrangement agent, effects pipeline) deployed on LangGraph Platform; VC demo driver; pro DJ testers.
CLaMP3 semantic search: Sub-100ms text/audio/hybrid search over 10K+ files; dual deployment (local vs. GCP GPU) with 15–30× speedup via batching/model reuse.
Web platform/payments: Built rent-to-own/perpetual payments (NestJS/React/Chargebee) in ~2 months; backward-compatible with legacy subs; zero production billing issues; discount engine with precedence rules and price floors.
Team: Interviewer/onboarding; established patterns for cross-platform code, rollout controls, and testing across desktop and services.

Pearse Audio Consultancy — Director/Consultant (2022–2023) Contracted by Density for DSP architecture, real-time audio processing, and software design on the DAW.

University of Portsmouth, School of Creative Technologies — Senior Lecturer, Computational Sound (2011–2020) Led music computing research group and MSc program; taught C++ audio development and spatial audio; supervised PhDs. Built EU-funded Compose with Sounds pedagogical DAW (solo dev, 7 years): custom C++/Qt engine, 10 languages, Mac/Win/Linux deployment in European schools; ADC 2019 talk and SMC 2019 paper.

Selected Projects

Density Copilot: AI voice assistant for music production; sole owner of Electron/React/ElevenLabs frontend; heavy contributor to LangGraph multi-agent backend; DensityMark DSL + streaming parser; Ableton Live integration; key fundraising differentiator.
Density DAW: Pro DJ DAW; cross-platform port; reliability/observability; real-time DSP; hardware integration; ~2,500 DAU.
CLaMP3 Audio Semantic Search: Research-to-prod ML pipeline in <1 month; MERT→CLaMP3 embeddings; FAISS; dual deployment; sub-100ms queries.
Density Web Platform (RTO/Payments): Full-stack licensing shift (one-time and $15×14 rent-to-own); Chargebee webhooks with locking; React flows; zero billing incidents.
Compose with Sounds: EU Commission–funded educational DAW; 10-language i18n; cross-platform; curriculum-aligned materials.

Education

PhD, Music Technology — University of Sheffield (2017) Thesis: “Agent-Based Graphic Sound Synthesis and Acousmatic Composition”
BMus — University of Sheffield (2010)

Publications & Talks (selected)

“No tracks or graphs? Designing sound-based educational audio workstations” — Audio Developer Conference (ADC), 2019 (talk).
“Composing with Sounds: Designing an Object-Oriented DAW for the teaching of Sound-Based Composition” — Sound and Music Computing Conference, 2019.
“An Agent-Based Approach to Interaction and Composition” — ICMC, 2014.