Founding CTO of AI and Robotics

5 percent equity

Write something impressive in your application before you press enter, and include 345 so I know you actually read this post. Add 465 at the very end of your application text.

About Orb

We are building Bloo, an ambient companion robot with eyes, ears, motion, and continuous memory. Phones and speakers cannot live with you in the same way. Bloo can. Our mission is to bring real presence and intelligence into everyday life through an AI native device that senses, understands, and adapts continuously.

Role

Own and build the full AI and software stack that powers Bloo. You design and code the core systems, make hard tradeoffs, and lead execution from edge to cloud. You are the third person at Orb with clear authority over AI, software, and the software parts of the hardware stack. You will recruit and mentor the team while shipping code every week.

Scope of authority

• Product and technical roadmap for AI and software
• Model and data strategy across edge and cloud
• Security, privacy, and safety for data and systems
• Reliability, observability, and release management
• Team building across platform, perception, and application layers

What you will do

Strategy and architecture that you also implement

• Define and build the system that links voice, vision, motion, and behavior in real time on Raspberry Pi 5 with a cloud brain for heavy lift tasks
• Choose models, toolchains, and runtimes and set the bar for latency, accuracy, and memory use
• Create the security model including threat model, authentication and authorization, key management, and data retention rules

Hands on model work

• Fine tune or train small models on top of strong foundations for speech, language, intent, and affect
• Apply LoRA and adapters, perform distillation to smaller targets, and quantize to run on device
• Build evaluation suites, guardrails, safety checks, and continuous regression tests for models

Edge intelligence

• Implement the local AI stack with Whisper cpp or Vosk, TinyLlama or Mistral GGUF, Coqui TTS or Mimic 3, and Mediapipe or OpenCV
• Deliver face detection, recognition, and basic affect cues
• Control servos for head and body and drive LEDs and the display for expressive output
• Keep full round trip interaction under two seconds from wake word to speech with streaming TTS, VAD, and careful buffering

Cloud and platform

• Stand up a service layer with FastAPI and gRPC for orchestration, memory, and personalization
• Design data pipelines for embeddings, session state, and long term memory with clear privacy controls
• Build OTA updates for firmware and models with signed releases and safe rollback
• Add metrics, tracing, and crash reporting for every module

Security and privacy

• Enforce least privilege across services and devices
• Implement secure boot, signed binaries, secrets rotation, and encrypted storage
• Write incident response runbooks and run live fire tests

Hardware software integration

• Work with mechanical and electrical design to shape compute, sensors, and power choices
• Define firmware interfaces for motors, microphones, speakers, and the display
• Own end to end performance and thermal budgets for real workloads

Hiring and leadership while coding

• Hire senior ICs across perception, platform, and application layers
• Create code standards, review culture, and a bias for shipping tested systems
• Stay in the codebase and model high quality engineering

Stack

• Python and C plus plus on device
• Whisper cpp or Vosk
• TinyLlama or Mistral GGUF
• Coqui TTS or Mimic 3
• Mediapipe and OpenCV
• FastAPI and gRPC
• Postgres and Redis
• Raspberry Pi 5 and GPIO
• Linux, containerization, and CI
• ROS 2 or lightweight pub sub like MQTT for module messaging

Nice to have

• CUDA or Triton, ONNX Runtime, Rust for safety critical parts, secure boot, TPM, Yocto, k8s where it helps

Compensation

• Salary
• 5 percent equity vesting over four years with a one year cliff
• Full time path after MVP with higher salary

How to apply

Show prior work that proves you can ship real time AI on device. A repo, a demo video, or a short tech brief is perfect. Tell me about one system you built that got faster and safer over time and what you cut to make the date. Put 345 in your note so I know you read this. Add 465 at the very end of your application text.