Skip to main content

Acoustics, Vibration & Audio Engineer · Creative Coder

Sergio Peña

A decade at the intersection of physics, perception, and product.
Meta · Apple · Google · Amazon · Boeing · Northrop Grumman

// about

About Sergio Peña

Deep expertise across acoustics, vibration, audio, and simulation, strongest where these disciplines intersect with product engineering and real-world physics.

Over a decade I've built a cross-industry body of work spanning consumer electronics, commercial aerospace, defense, and environmental noise control engineering. Based in the San Francisco Bay Area. From nuclear submarines to iPhones, from wide-body aircraft to smart glasses.

I combine rigorous testing, multiphysics modeling, DSP, and data-driven analysis. I automate what should be automated, measure what others overlook, and build the tools I wish existed. Lately, agentic AI development workflows let me ship lab tooling in hours that would have taken days — measurement utilities, automation harnesses, validation pipelines that close gaps and save hours per test cycle.

The maker mindset runs deep. GeoNoise brings ISO 9613-2 noise modeling to the browser with WebGPU compute and real-time dynamic scenarios. ImpedanceTuber modernizes ISO 10534-2 impedance-tube measurement into a web app with 75 documented physics equations. SonoTag does zero-shot sound detection with OpenFLAM (Foundation Language-Audio Model) and free-form text prompts.

Education
  • Master, COMSOL Multiphysics - U. de Málaga (99/100)
  • Graduate Studies, Aerospace Science - UPC Barcelona
  • B.Eng. Acoustics & Audio - U. Tecnológica de Chile
Continuing Ed
  • MITx - Deep Learning, Material Mechanics (3-part), Probability
  • IBM - Python, PyTorch for ML
  • LMU Munich - FEM / Numerical Methods in Python
  • AltaSim - Acoustic COMSOL Multiphysics
  • MathWorks - ML, Deep Learning, DSP, ODEs, Stats
  • Audinate - Dante Certification Level 1
  • Listen, Inc. - SoundCheck (Beginner & Advanced, 2026)

Beyond the lab: live FOH mixing, FPV drones (full acro), Sunfish sailing, scuba, biking, guitar, flute - and usually some new system to learn every month. From the outside it looks scattered. From the inside it's the same instinct on repeat: learn the physics, develop feel, build fast, ship something, repeat.

// signal_path

Signal Path

From environmental consulting to defense submarines to wide-body aircraft to consumer devices. A decade of making things sound right.

// filter by domain

Meta

Acoustic Engineer

Nov 2025 - Present Consumer Electronics
  • Reality Labs, Audio Experience Team — Wearables Audio Tuning Team (WATT); Ray-Ban smart-glasses acoustic tuning
  • Drove speech-quality KPI assessment for shippable Ray-Ban tunings using HEAD acoustics ACQUA (POLQA, PESQ, 3QUEST)
  • Refactored the 360° spatial-audio measurement CLI (turntable, stimulus, mic array): eliminated manual user inputs to remove operator error and added full paper trail of test conditions
  • Built a mic-health validation tool that pre-screens channels and catches frame drops and dead mics
  • Architecting a full-stack device control center (React + FastAPI) with visual workflow editor for codeless test automation
  • Shipped a fleet portal and database covering 19 product families (QR scanning, analytics, Brother label printing)
Sunnyvale, CA contract via LTTS

Apple

Electronic Noise / Mic Co-Existence Engineer

Feb 2025 - Nov 2025 Consumer Electronics
  • Owned eNoise spec compliance for next-generation Apple devices (iPhone 17 Pro, M5 MacBook Pro line); held the acceptance bar against the team's hard acoustic criteria, gating launch readiness
  • Drove cross-functional root-cause and remediation with mechanical, electrical, and product-development partners when units failed spec
  • Architected automated acoustic test suites for spec compliance
  • Built BASH / Python / MATLAB / JMP scripts for measurement processing
Cupertino, CA contract via CPrime

Universidad de Málaga

Master in COMSOL Multiphysics: Structural Mechanics & Acoustics — 99/100

2024 - 2025 Education
  • 60 ECTS — Multiphysics FEA: structural mechanics, acoustics, CFD, MEMs, optimization, MATLAB LiveLink
  • Thesis: Transient thermoviscous acoustics and FFT-based impedance extraction in actively occluded micro-perforated panels. Advisors: R. Pico (UPV), B. Ivorra (UCM)
Málaga, Spain (remote) degree

Northrop Grumman

Principal Acoustic Engineer

Mar 2024 - Feb 2025 Defense
  • Led acoustics/DSP efforts for Marine Systems
  • Applied SEA models to reduce SSN(X) equipment noise
  • Analyzed CVN-80 rotational vibration data and developed test viability strategy for leadership and the Navy
Sunnyvale, CA full-time

Google

Acoustic Test Engineer

Jan 2024 - Mar 2024 Consumer Electronics
  • Conducted anechoic playback-quality testing and UX listening studies for Pixel 9
  • Accelerated lab throughput by reorganizing inventories
  • Assisted display team with root-cause analysis of early-field failures
Mountain View, CA contract via YOH

Apple

Electronic Noise Engineer

Jan 2023 - Dec 2023 Consumer Electronics
  • Investigated eNoise across MacBook, iPad, Mac Mini
  • Executed Laser Doppler Vibrometry and acoustic-camera testing for failure analysis
  • Led eNoise efforts on the 2023 M3 MacBook Pro architecture
  • Pioneered autoencoder-based 1D measurement anomaly detection
Cupertino, CA contract via CPrime

Amazon Lab126

Audio Algorithms SQA

Oct 2020 - Dec 2022 Audio Software
  • Evaluated wake-word, AEC, VoIP, and playback quality for Alexa devices
  • Developed sequential Bayesian testing to reduce test time by 30-60%
  • Automated SoundCheck suites via Python/MATLAB
  • Developed vision-based convolutional neural network model for robust barge-in detection
Sunnyvale, CA full-time

Boeing

Acoustic Engineer

Jan 2019 - Sept 2020 Commercial Aerospace
  • Led cross-functional teams verifying subsystem noise requirements for 777X
  • Built MATLAB ray-tracing tools predicting ramp noise
  • Modeled complex wave diffraction using VA-One for noise prediction
  • Delivered field-data apps built in MATLAB to engineers
Everett, WA full-time

Universitat Politècnica de Catalunya

Graduate Studies, Aerospace Science & Technology

2017 - 2018 Education
  • All coursework completed (60 ECTS) — supersonic shocks, flight dynamics, astrodynamics, space systems, ECLSS. Thesis deferred for Boeing 777X position.
Barcelona, Spain coursework

Vibro-Acoustic Consultants

Acoustic & Vibration Engineer

Apr 2016 - Jun 2017 Environmental Consulting
  • Managed environmental noise and vibration compliance for academic, medical, and residential projects
  • Deployed and maintained in-house remote monitoring stations
  • Monitored noise in animal research facilities to maintain welfare-driven thresholds
  • Oversaw continuous noise monitoring during VA Hospital expansion
San Francisco, CA full-time

Pasto Seco / Los Piures de Tuman

Sound Engineer

2014 - 2017 Live Audio
  • FOH mixing and PA sound at venues across the Bay Area and Chile
  • Stages: Lollapalooza Chile 2015, Ballantine's Summer Tour 2015, SurfBeats 2015, Carnaval San Francisco 2017, The Starry Plough, Moe's Alley, Brick and Mortar, The New Parrish
San Francisco / Santiago, Chile freelance

Maria Magdalena Music

Stage Manager

2014 - 2014 Live Audio
  • Coordination of transportation, stage requirements, and staff movements
  • Stages: CreamFields Chile 2014 (Main Stage), Mystery Land Chile 2014 (MicNest Stage), La Roma (Teatro Italia), LoveParade (Paseo Bulnes), Club Chocolate (Main Stage)
Santiago, Chile freelance

Universidad Tecnológica de Chile

B.Eng. Acoustics & Audio Engineering

2010 - 2014 Education
  • Acoustics: architectural design, environmental noise, vibration control, noise isolation, underwater acoustics, ultrasound
  • Audio: electroacoustics, DSP, loudspeaker design (Thiele-Small), live sound, recording, C++
Santiago, Chile degree

Hurtado y Asociados

Chief Audio-Visual Technician

Jan 2012 - Aug 2013 Live Audio
  • Led a team of sound technicians for mid- and large-size events
  • Audiovisual installations, equipment operation, and production coordination
  • Responsible for transportation logistics and equipment storage
Santiago, Chile full-time
// skills

Skills & Tools

Expertise across the full stack, from anechoic chambers to neural networks.

Acoustics & Vibration

  • Acoustic & vibration testing
  • eNoise / mic coexistence
  • Environmental noise monitoring
  • Psychoacoustics & NVH
  • ISO / IEC / ITU standards

Simulation & Modeling

  • COMSOL Multiphysics
  • Wave6 / VA One
  • SEA / FEA / BEM
  • ISO 9613-2 propagation
  • Ray tracing

Signal Processing & DSP

  • Digital signal processing
  • FFT / spectral analysis
  • Beamforming & AEC
  • Filter design
  • Audio quality metrics (POLQA, PESQ, 3QUEST)

Software & Data

  • Python
  • MATLAB
  • TypeScript / JavaScript
  • PyTorch / ML for data analysis
  • JMP Pro / statistical analysis
  • Git, lab automation

Hardware & Lab

  • B&K / G.R.A.S. microphones
  • HEAD acoustics ACQUA
  • LMS Test.Lab / SoundCheck
  • Laser Doppler Vibrometry
  • Tektronix oscilloscopes, Chroma EE load
  • Arduino / Raspberry Pi / SparkFun

Live Audio

  • FOH mixing
  • System rigging
  • Dante networking
  • Live recording & mixing
  • Remote monitoring stations

AI-Assisted Engineering

  • Agentic AI development workflows
  • Multi-perspective code review methodology
  • Lab-tool prototyping with AI assistance
  • Anthropic SDK / Claude Code
  • Local LLM serving (vLLM, Ollama)
// projects

Projects

GeoNoise

Browser-based environmental noise modeling

Full-stack noise propagation simulator implementing ISO 9613-2 and CNOSSOS-EU with dual CPU/WebGPU compute backends. Coherent two-ray phasor model on WGSL compute shaders, supporting real-time dynamic scenarios (moving sources and receivers) beyond what static batch-prediction tools deliver. Built as a TypeScript monorepo with 7 packages and 1,500+ physics validation tests; Stripe-powered monetization.

  • ISO 9613-2 and CNOSSOS-EU compliant propagation engine
  • WebGPU compute shaders for real-time noise maps
  • Real-time dynamic scenarios: moving sources and receivers
  • 1,500+ tests with physics validation against analytical solutions
  • Dual CPU/GPU backends with automatic fallback
  • Monetization: Stripe Checkout + email-verified activation
  • Free browser alternative to SoundPLAN and CadnaA
TypeScriptWebGPU (WGSL)Canvas 2DWeb WorkersStripeVercelVitest
Visit →

ImpedanceTuber

Modern web ISO 10534-2 impedance-tube measurement system

Modernization of a 2005 BERSALFA MATLAB GUI into a browser-based measurement system. FastAPI Python physics engine + Next.js frontend with 75 documented physics equations (full ISO vs. MATLAB vs. Python audit), H1 / H2 / Hv transfer-function estimators, viscothermal corrections, switching-technique calibration, and A–F quality grading. Architecture complete; parked until I can collect more real-tube measurement data to validate end-to-end.

  • ISO 10534-2 compliant impedance-tube measurement
  • 75 documented physics equations (ISO/MATLAB/Python audit)
  • H1 / H2 / Hv transfer-function estimators
  • Viscothermal corrections and switching-technique calibration
  • A–F quality grading for each measurement
  • Pure-physics implementation, no MATLAB runtime dependency
PythonFastAPINext.js 14NumPy / SciPyTypeScript

SonoTag

Real-time audio tagging with zero-shot AI inference

A browser-first audio console for live microphone input, real-time spectrogram visualization, and zero-shot sound event detection using OpenFLAM (Frame-Wise Language-Audio Modeling, Wu et al., ICML 2025). Define any sound category via free-form text prompts, no pre-training required. Features live YouTube video analysis, cumulative statistics with PDF/CDF histograms, and a glassmorphism UI.

  • Zero-shot sound detection: describe any sound, FLAM detects it
  • Real-time spectrogram with adjustable frequency range
  • Live YouTube video analysis in floating draggable modal
  • Cumulative statistics with PDF/CDF/Histogram charts
  • System-aware buffer recommendations based on hardware
  • Webcam + mic simultaneous capture
TypeScriptReactViteFastAPIPythonOpenFLAMWeb Audio API
Visit →

vibration-pdm

Edge ML for noise / vibration predictive maintenance

Research project exploring how raw acoustic and vibration sensor streams can drive fault detection and condition monitoring directly on industrial equipment in the field, without cloud round-trips. Edge-deployed ML stack; currently in pre-training and architecture research.

  • Edge-deployed inference, no cloud round-trips
  • Multi-sensor fusion: acoustic + vibration streams
  • Targeting fault detection and condition monitoring
  • Currently in pre-training and architecture research
PythonPyTorchEdge MLASUS GB10

acoustics-rag

On-demand reference over an acoustics / engineering textbook corpus

Personal retrieval-augmented reference tool for noise / vibration consulting. Local-LLM backend over a curated engineering-textbook corpus, queryable through a Telegram interface. Standards-grounded reference answers from the field, daily-driver tool.

  • Local-LLM backend, no API costs or cloud dependency
  • Curated acoustics / engineering textbook corpus
  • Telegram interface for query-from-anywhere access
  • Standards-grounded responses for in-field consulting
PythonLocal LLM (vLLM)ChromaDBTelegram Bot API

Cellcraft

Browser-based ecosystem simulation with genetics

An ecosystem simulation where cells, predators, protectors, and plague compete for survival. Entities have heritable genetics with mutation, seasonal food cycles drive population dynamics, and an SIR epidemic model spreads disease through contact. Built as a Turborepo monorepo with pure simulation logic separated from rendering.

  • Heritable genetics with trait mutation across generations
  • SIR epidemic model: disease spreads through entity contact
  • Seasonal food cycles driving population booms and busts
  • 4 entity types: cells, predators, protectors, plague
  • Pure simulation engine, zero DOM dependencies
  • Multiplayer sync layer via Pusher (in progress)
TypeScriptNext.jsCanvas 2DTurborepoPusher (multiplayer)
Visit →

* This very page is also a project. The waveform above reacts to your cursor, clicks spawn parasitic harmonics, and the badges in the nav are a working oscillator. Oh, and there's a mic icon up there that makes it react to your voice. Built with Astro, Canvas 2D, and zero frameworks. Reviewed by 13 AI Elders who argued about it for 3 rounds.

// contact

Let's Talk

Need someone who speaks both COMSOL and Python? Want to push browser-based acoustics further? Or just want to argue about mic placement at FOH? Available for contract and consulting engagements.

↓ CV View & print in LinkedIn Connect with me