Founding Engineer
Location: Amsterdam / Netherlands preferred (hybrid-flex) Type: Founding engineer (full-time) Start: ASAP
Build TILDR. Turn any on-screen text into a comprehension‑boosting listening experience. You’ll co‑own the architecture, ship an alpha in 60–90 days, and set our engineering culture.
About a10n
a10n (augmentation) builds humane speech tech. We stand on 5+ years of customizable speech technology experience at DAISYS –hard‑won know‑how in controllable prosody and voice disentanglement which we’re rebuilding from first principles at a10n to power TILDR, our intelligent listening app with personalized voices, adaptive pacing, importance cues, and voice interaction.
Why now / Unfair advantage
- 5+ years of controllability‑first TTS R&D.
- Experienced founding team + connections to make this startup fly.
- A “simple‑first, rebuild‑right” approach: fresh codebase, modern toolchain, fast shipping.
Our values & culture
- Build > talk: ship small, ship often, learn fast.
- High bar, low ego: strong opinions, loosely held; crisp code reviews.
- Curiosity as a practice: hard problems are research quests, not blockers.
- Can‑do ownership: bias to action, own outcomes end‑to‑end.
- Fun & craft: playful details, micro‑interactions, great audio taste.
- User delight + measurable impact: wow moments and numbers to prove it.
- Privacy by default: dignity for the end‑user is non‑negotiable.
The role (you’ll own v1 → v1.x)
- Co‑found the company with product + technical leadership.
- Architect and ship the TILDR prototype & MVP quickly, then iterate to reliability and scale.
- Use AI as a force multiplier (codegen, testing, eval loops) while maintaining high quality.
- Make pragmatic build/buy choices across TTS, LLMs, accessibility APIs, and desktop frameworks.
- Establish the core engineering culture: velocity, craftsmanship, and user‑obsession.
Leadership & managerial growth (learn‑by‑doing)
- No prior management required: willingness to learn.
- Lead small efforts now, grow into team lead after MVP traction.
- Run crisp code reviews, set lightweight processes (weeklies, demo day, release rhythm).
- Help define the engineering ladder and hiring rubric.
What you’ll do
- Ship the MVP (first 60–90 days):
- Desktop shell: Mac first (Windows later) for on‑screen text capture & direct text access via accessibility APIs and browser integrations.
- Real‑time TTS pipeline: controllable prosody (rate, emphasis, pauses, “importance cues”).
- Voice navigation: natural language commands and quick comprehension checks.
- Local perf: caching, streaming, low‑latency audio playback; privacy‑respectful data handling.
- Python services: orchestrate TTS/LLM flows (FastAPI/asyncio/WebSockets), evaluation loops, and prompt/tooling pipelines.
- Own the stack: architecture, CI/CD, telemetry, error budgets, and release cadence.
- Prototype fast: build scrappy experiments, instrument with metrics, and cut what doesn’t move the needle.
- Hire and lead: help grow a small, senior‑lean team after MVP traction.
- Create the wow: design signature interactions and land them with polish.
Our (initial) tech preferences
- App: Tauri (preferred) or Electron – Rust core + TypeScript/React UI.
- Core services (Python‑first): Python 3.10+ (FastAPI/uvicorn, asyncio, pydantic), prompt/LLM orchestration, evaluation harnesses.
- Systems / back end: Rust or Go services; REST/WebSockets/IPC; Postgres/SQLite.
- TTS & LLMs: neural TTS with SSML/IPA controls; promptable LLMs for parsing/QA; selective on‑device inference.
- Bridging: IPC (local HTTP/WS), or FFI via pyo3/maturin when needed.
- Tooling: GitHub, CI, typed code (mypy/ruff), pytest/property tests, observability by default.
Self‑qualify scorecard (read this first)
- You’ve shipped a desktop app or dev tool end‑to‑end.
- You move from fuzzy problem → measurable outcome fast.
- You use AI tools to multiply speed and quality.
- You’re fluent in TypeScript + Python, and comfortable with one systems language (Rust or Go).
- You care about latency, audio pipelines, and clean architecture.
Must‑haves
- Master’s (or higher) in Computer Science or closely related field, or equivalent demonstrable expertise.
- Evidence of shipping: production desktop or developer tools, maintained over time.
- Strong with TypeScript and Python (3.10+): concurrency (asyncio), API design (FastAPI), testing.
- Comfortable with Rust or Go for low‑latency components and the desktop core.
- Experience integrating TTS/LLM services and handling real‑time audio streams.
- Comfort with accessibility APIs, browser‑extension bridges, or text‑ingestion pipelines.
- Bias to action, strong product taste, and a builder mindset that embraces AI tools.
- Excellent communication; you turn fuzzy problems into shipped product.
- Early‑career, high‑agency profile (e.g., 0–5 years post‑Master’s) welcomed.
Nice‑to‑haves
- Rust in production (Tauri) or deep Electron experience.
- Prosody/voice controls (SSML, phonemes/IPA), audio DSP basics, or latency tuning.
- C++ for audio/DSP or legacy SDK integration when needed.
- On‑device inference (ggml/gguf, WebGPU, CoreML) and model evaluation chops.
- FFI experience (pyo3/maturin) and packaging discipline (Poetry/uv).
- Security/privacy by design; GDPR familiarity.
Compensation
- Founding‑level equity (target 2.5–5.5%, seniority‑dependent) + salary band €60–90k pre‑seed (adjusts post‑seed).
- Standard 4‑year vesting / 1‑year cliff. We’ll tailor the mix for the right co‑founder.
How we work
- User‑centric experiments, short cycles, honest metrics.
- “Simple first” architectures that survive success.
- High standards, low ego, fast feedback.
- Weekly demo day to celebrate wins, surface risks, and keep craft high.
What success looks like in 90 days
- Alpha Mac build capturing selected text and streaming TTS with “slower/faster/deeper” commands.
- Stable sub‑300 ms perceived latency for short utterances on common hardware.
- Robust crash/error telemetry and a weekly release cadence.
- A measurable comprehension lift vs. plain TTS on a small internal benchmark.
- At least one “wow‑moment” demo that makes evaluators say “I want this,” plus 3+ signature interactions with >70% user delight scores.
Hiring process (fast and transparent)
- 25–30 min intro call (fit, product taste).
- 60 min technical deep dive + code walkthrough (bring a repo you’ve shipped).
- 2–3 h live working session or take‑home: text‑capture → streaming TTS spike in Tauri (or your stack).
- Wow review: propose one delightful micro‑interaction for TILDR.
- Founder chat (culture & roadmap) + same‑day references.
- Decision within 72 hours of the final chat.
Location & work authorization
- Amsterdam / Netherlands preferred (hybrid‑flex). EU work authorization preferred; relocation considered and visa support possible for exceptional candidates.
How to apply
Send a short note, links to things you’ve shipped, and a CV to hello@a10n.io - or share a weekend‑scale demo (e.g., a minimal desktop app that captures selected text and streams it to a TTS service with a “slower/faster” control).