System design of complete system for predictive maintenance, from IoT sensors and gateway, to data processing system and Machine Learning, web service with open API and web application frontend. 

Machine Learning pipeline in Python with PyData stack.
Backend development with Flask, PostgreSQL/TimescaleDB, RabbitMQ.

Firmware development on STM32 and NRF52 based sensors. CAT LTE 4G and Bluetooth wireless communication.

Head of Data Science & CTO

Development of the Flowhub.io visual-programming IDE, and associated programming runtimes for web-systems, image processing and microcontrollers.
Participating in EU research projects as part of Horizon 2020.

Creating and maintaining the image processing pipeline: http://imgflo.org
Overseeing work on image analytics / feature extraction.

Created scalable heterogeneous message-passing system used in backend: https://github.com/msgflo/msgflo
Created solution for automatic scaling of compute resources on Heroku. https://github.com/flowhub/guv

Built initial version of hybrid/native Android app, onboarded dedicated developer.

Work on http://flowhub.io visual programming IDE and tooling
Node.js server-side work in JavaScript/CoffeeScript

Data Engineer

Software development of internal software, mostly C++ based, used in microphone array systems.

Bringup and maintenance of an embedded Linux distribution based on Ubuntu LTS.
Bringup of Linux server platform using CentOS/RHEL.

Improving QA processes through the use of Jenkins for continuous integration&delivery, mandating code reviews with git.

Software Engineer

Software development of open source software targeting GNU/Linux systems for customers, including Nokia and Canonical.

Software Developer

Software Development on Maliit, the text input method and virtual keyboard solution for Nokia N9 and MeeGo.

External Software Developer

Have developed most kinds of software in Python. Web backend, Data Science and Machine Learning pipelines, scripting and practical automation, automated tests for software and hardware, desktop GUI application development.

Have developed and deployed Machine Learning systems for multiple companies. 
Extensive experience in time-series analysis, including time-series classification, clustering, Event Detection and Anomaly Detection.
Expert in Machine Learning applied to Audio, including Audio Classification, Sound Event Detection, Audio Quality Metrics. Good foundation also in speech and music tasks.
Expert in creating small, computationally efficient models for execution in constrained environments such as embedded devices and microcontrollers.
Maintainer of the emlearn TinyML inference engine on Github.

Have developed several applications utilizing signal processing. Examples including an image processing web service that applied artistic filters, rescaled and optimized images for delivery on web. It served thousands of websites over the years. At Soundsensing I wrote the entire audio processing pipeline for acoustical sensors. Also used to be a contributor to the GIMP and MyPaint open-source graphical software, and have presented at Libre Graphics Meeting et.c.

Have worked extensively with both microcontrollers and Linux based embedded devices, doing both application development, system design, and hardware bringup. Done the full development and deployment cycle for IoT devices based on STM32, ESP32 and Raspberry Pi.

Have deployed and maintained production-grade systems on Linux, both on IoT/embedded systems, web services and mobile devices. Also used Linux exclusively since 2008 on my own PCs and servers.

Was part of the Meego team at Nokia, their last Linux phone. Developed the virtual-keyboard solution (codename Maliit).

Developed an Ubuntu-based embedded device platform for microphone arrays, for Squarehead Technology.

Built and maintained Raspberry Pi-based IoT door system for municipality of Oslo.

Have developed, deployed and maintained several production web services on Linux.

Have maintained some 30 different packages for Arch Linux, in the Arch Linux User repository.

Have designed, developed and maintained web services for multiple companies. These services have been in operation for many years and processed billions of requests.

I have developed and maintained open-source software since 2010. Projects I have been involved with include GNOME, GIMP, MyPaint, Flowhub.io, imgflo, microflo and emlearn.

I have advised several companies on stategic publishing of open-source, including Nokia, TheGrid, Flowhub and Soundsensing.

Machine Learning Engineer for 5 years, Python developer for 12+ years

Jon Nordby

Machine learning for microcontroller and embedded systems. Train in Python, then do inference on any device with a C99 compiler. Open-source under MIT license.

emlearn

imgflo is an image-processing runtime built on top of GEGL. Image processing pipelines can be visually programmed using node-based / flow-based-programming, using the Flowhub.io IDE.

imgflo

An IoT doorsystem controller for standard electronic door strikes and door opening actuators. Coordinates the physical opener buttons, mode switches, and offers a HTTP/MQTT interface for unlocking using a mobile app. Focus on ease of prototyping and adapting to changes.
Used in first stage of the "Oslonøkkelen" project, to enable citizens extended access to municipality services.

UnlockOslo

Implementation of environmental sound classification on STM32 microcontroller. Efficient CNN models trained with Keras/Tensorflow. State of the art on the Urbansound8k at the time.

Personal details

Jon N.

Summary

Other

Technical skills

Work Experience

Education

Personal Projects