Norman Heckscher
Coal Geologist at QGESS
Moranbah, Queensland, Australia
Summary
🤩
Rockstar🎓
Top SchoolNorman Heckscher is an experienced coal geologist based in Moranbah, Queensland, with over a decade of hands-on experience across mining, coal testing, and metallurgical operations. He combines field and lab expertise from roles at QGESS, Bureau Veritas and Anglo American with a solid academic foundation in geology, mathematics, computing and electrical engineering from the University of Tasmania. Beyond traditional geology he has a notable interest and practical contributions in machine learning and TensorFlow—refactoring and modernizing several high-profile examples and RNN language-model codebases—which speaks to strong numerical, coding and model-training skills uncommon in his discipline. Known for improving reproducibility and performance (TensorBoard logging, beam search, GPU memory fixes), he bridges domain knowledge and low-level engineering to solve data- and model-driven problems in resource evaluation. Colleagues value his pragmatic, multidisciplinary approach: grounded field judgment married to software-savvy optimization.
10 years of coding experience
1 year of employment as a software developer
Graduate Diploma of Science, Earth Science, Geology/Geophysics, Graduate Diploma of Science, Earth Science, Geology/Geophysics at University of Tasmania
Stackoverflow
911reputation
504kreached
12answers
0questions
Github Skills (25)
beam-search10
python10
artificial-neural-networks10
machine-learning10
recurrent-neural-networks10
deeplearning-ai10
deep-learning10
tensorflow10
neural-network10
nlp10
n9
rnn-model9
tensorboard9
mask-rcnn9
faster-rcnn9
Programming languages (3)
Github contributions (5)
hunkim/word-rnn-tensorflow
Jan 2017 - Apr 2017
Multi-layer Recurrent Neural Networks (LSTM, RNN) for word-level language models in Python using TensorFlow.
Role in this project:
ML Engineer
ML EngineerContributions:50 commits, 16 PRs, 31 pushes in 3 months
Contributions summary:Norman made significant contributions to the `word-rnn-tensorflow` repository, focusing on improving the training and sampling processes for a word-level language model. Their work included adding features like TensorBoard logging, beam search sampling, and optimizing the code for better performance by moving operations within the TensorFlow graph. The user also addressed bugs related to batch pointers, epochs, and GPU memory allocation. Finally, the user refactored and cleaned up code, demonstrating a strong understanding of the project's underlying machine learning concepts.
pythonmulti-layerctcword2vectensorflow
google-deepmind/learning-to-learn
Jan 2017 - Feb 2017
Learning to Learn in TensorFlow
Role in this project:ML Engineer
ML EngineerContributions:5 commits, 2 PRs, 3 comments in 16 days
Contributions summary:Norman primarily contributed to refactoring and updating the code base to be compatible with newer TensorFlow versions (1.0 and release candidates). They modified code related to batch normalization, recurrent neural networks, and general module structures, reflecting a focus on model architecture and training procedures. The contributions also included adjustments to file paths and dependencies related to the dataset and build configuration within the project's problem-solving modules.
deep-learningneural-networkslearning-to-learntensorflowartificial-intelligence
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.