Johan Eklöf
Chief Coffee Brewing Officer at salbo.ai
Sweden
Summary
👤
Senior🎓
Top SchoolJohan Eklöf is a seasoned engineering leader and founder with 16 years of experience building high-impact systems across sustainability reporting, conversational AI, online advertising, and large-scale gaming. As Head of Engineering at Normative and now leading salbo.ai, he combines system-architecture rigor with hands-on ML work—recently developing an unsupervised deep model for chat agents that matched ConvAI2 top performers. He has a strong track record of turning analytical insights into business wins, from beating competitors on ad profit optimization to delivering procured software for the Swedish military on schedule. An active open-source contributor in the Julia ML and data ecosystems (notably improving tensor ops in Flux.jl and shape inference in TensorFlow.jl), he brings deep numerical and systems expertise. Johan’s background in engineering physics and scientific computing powers a pragmatic approach to performance and correctness that shows up in production and research alike. Outside work he styles himself Chief Coffee Brewing Officer, hinting at a practical, detail-oriented personality that fuels both experimentation and delivery.
16 years of coding experience
12 years of employment as a software developer
Software Engineering, Software Engineering at Tongji UniversityMaster of Science, Engineering Physics with a minor in Scientific Computing, Master of Science, Engineering Physics with a minor in Scientific Computing at Uppsala University
English, Swedish
Github Skills (29)
fonts10
shader10
freetype10
testing10
machine-learning10
rnn-model10
inference10
glsl10
c1110
n10
data-structure10
c1710
deep-learning10
tensorflow10
text-rendering10
Programming languages (6)
Github contributions (5)
malmaud/TensorFlow.jl
Mar 2018 - Jun 2018
A Julia wrapper for TensorFlow
Role in this project:
ML Engineer
ML EngineerContributions:21 commits, 12 PRs, 20 comments in 2 months
Contributions summary:Johan primarily focused on extending the shape inference capabilities of the TensorFlow.jl wrapper. They added shape inference for operators like `Select`, `ArgMax`, and `ArgMin`, and implemented shape inference for binary operations. The user also improved the dynamic length RNN implementation and added tests. Additionally, they addressed issues related to saving and restoring TensorFlow models.
julia-wrappermlmachine-learningtensorflowjulia
JuliaData/DataFrames.jl
Aug 2016 - Sep 2016
In-memory tabular data in Julia
Role in this project:Back-end Developer & QA Engineer
Back-end Developer & QA EngineerContributions:1 review, 5 commits, 25 PRs in 1 month
Contributions summary:Johan made several contributions focused on improving the stability and correctness of the `dataframes.jl` library. These include fixing bugs related to grouping operations, specifically handling empty data frames and preventing crashes. They also addressed issues with type handling during concatenation and implemented optimizations to reduce redundant sorting operations. Furthermore, the user wrote and updated tests to ensure that the library's functionality works as intended.
memorydataframesdatatabular-datadata-frame
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.