🤘 I don't always listen to metal, but when I do my neighbours do too!
Late to the Party 🎉 is about insights into real-world AI without the hype.
Hello internet,
Time flies when you're having fun. Fly to the bottom of this issue to find 5(!) ML positions at ECMWF! Until then, let's enjoy some machine learning!
The Latest Fashion
- A personal assistant and chatbot that works locally with your (e.g. Obisidan) notes?!
- Jupyter now has an %ai magic for generative AI
- Prompt-tools runs your prompt against different LLMs for you to pick and choose
Got this from a friend? Subscribe here!
My Current Obsession
My chatGPT course was featured by Skillshare as one of their most popular classes in July, which has just brought in 100 new students in 2 days, which is pretty incredible. I also got a few really kind reviews, which is always nice!
I’ve been reading up on electric work a bit, as I’ve been obsessing about home automation a bit lately. So I’ll be working on never having to touch a light switch again this weekend.
Apart from that, currently, the first festival I ever went to, and used to go to for years, is on: Wacken. And they’re live-streaming the entirety of it. So while I’m old and out of touch, I can enjoy the bands and reminisce of the times! Right now, Killswitch Engage is playing.
This email is late because I have been playing Baldur's Gate 3. Sorry about that, but the game is just really good.
Thing I Like
I bought these Shelly relays for home automation. So once I’m done writing this, I’ll head out to the local hardware store and grab a bunch of electronics to install some fun automation in my home! If you don’t hear from me next week, you’ll have to assume I didn’t switch off the breaker… Let’s hope for the best!
Hot off the Press
I’ve been a bit quiet last week, but still managed to post a few things on socials. Linkedin liked the text visualization browser I shared with this newsletter 4 months ago. Mastodon was quite happy about getting recognition for their work.
In Case You Missed It
People have discovered the interpretability section of ML.recipes, so that’s great!
Machine Learning Insights
Last week I asked, What’s the difference between ML engineers and ML scientists?, and here’s the gist of it:
The roles of ML engineers and ML scientists may overlap, but they involve distinct responsibilities and areas of expertise within machine learning.
ML Engineers: ML engineers are focused on practically implementing and deploying machine learning models in real-world applications. Their main goal is to build and optimize machine learning systems that can be integrated into products and services. They are skilled in software development, data engineering, and machine learning frameworks.
Responsibilities of ML Engineers:
- Data Pipelines: ML engineers can clean and prepare data for training machine learning models, although there are also dedicated data engineers these days!
- Model Deployment: ML engineers deploy trained models into production environments, making them accessible to end-users or applications.
- Scaling and Optimization: They optimize the performance and efficiency of machine learning systems to handle large-scale data and real-time processing.
- Software Integration: ML engineers integrate machine learning capabilities into existing software systems or develop new applications.
Example from Meteorological Data: An ML engineer working with meteorological data might take a machine learning model to predict rainfall based on historical weather data. They would develop a pipeline for the data and deploy it as part of a weather forecasting application.
ML Scientists: ML scientists, on the other hand, are more focused on the research and theoretical aspects of machine learning. They delve into the development and improvement of algorithms and methodologies. They conduct experiments, analyze results, and contribute to advancing machine learning theory.
Responsibilities of ML Scientists:
- Algorithm Development: ML scientists work on creating novel machine learning algorithms or improving existing ones to solve complex problems.
- Research and Experimentation: They design experiments to test the effectiveness of different algorithms and approaches on various datasets.
- Performance Evaluation: ML scientists analyze the performance of machine learning models, including their strengths and limitations.
- Publications: They often publish their research findings in academic journals or present them at conferences to share knowledge with the community.
- Advancing the Field: ML scientists contribute to the theoretical foundations of machine learning, pushing the boundaries of what is possible with technology.
Example from Meteorological Data: An ML scientist working with meteorological data might explore novel neural network architectures or optimization algorithms to improve the accuracy and efficiency of weather prediction models. They would conduct experiments to evaluate the performance of these new methods and publish their findings to benefit the machine learning research community.
In summary, while ML engineers and ML scientists contribute to the development and application of machine learning, their roles differ in practical implementation versus research and theoretical advancements. ML engineers focus on building and deploying machine learning systems, while ML scientists concentrate on algorithm development and advancing the theoretical aspects of machine learning.
That’s why we’re hiring multiple of those positions at ECMWF. I posted them in the job corner below!
Data Stories
Creator burnout has been a huge topic for a while.
So Tasty Edits made a survey and compared high-earning and low-earning creators and their coping mechanisms.
Interesting to see that the feeling of stress is almost equal across the bank!
Question of the Week
- What is an auto-regressive model, and what applications do you see for them?
Post them on Mastodon and Tag me. I'd love to see what you come up with. Then I can include them in the next issue!
Job Corner
The ECMWF is hiring 5 people that touch machine learning right now!
Four positions in the core team to develop a data-driven machine learning weather forecasting:
- Machine Learning Engineer: Focus on model optimization and parallel implementations to train large machine learning models on vast datasets. Prior experience with deep learning frameworks, model optimization, and memory footprint improvements is essential. Background in earth-system modelling is welcomed.
- Observations and Data Assimilation Expert: Interface observations with machine learning algorithms and play a vital role in data assimilation. Exceptional interpersonal skills and expertise in using earth-system observations are highly valuable for this role.
- Machine Learning Scientist for Learning from Observations: Contribute to making future earth system predictions from observation data using deep learning frameworks. Experience in earth-system observation data and data assimilation approaches is desirable.
- Machine Learning Scientist for Precipitation: Specialize in accurate precipitation predictions with generative machine learning models. Experience with GANs, VAEs, or Diffusion approaches is advantageous, along with expertise in using neural networks for precipitation prediction.
And one on the EU project Destination Earth
You will leverage cutting-edge machine learning techniques and statistical methods to support uncertainty quantification for weather-induced extremes in the revolutionary Destination Earth (DestinE) Digital Twin. Your work will contribute to more accurate and reliable predictions, shaping the future of weather forecasting and its impact on climate understanding and resilience. If you're a proactive and talented individual with a passion for Earth System Science and a flair for machine learning, apply now and make a meaningful difference in tackling climate challenges.
(I like to stress that these positions are, as always, written by a committee, so if identify as part of an under-represented minority, please consider applying, even if you don’t hit every single bullet point. Also, salaries are posted, and they're "net of tax". Not the $900,000 Netflix recently posted, but certainly not the worst.)
Tidbits from the Web
- This Orca video made my day
- This ICML poster is rightfully going viral
- A Public Service Announcement about coffee
No neighbours were harmed in the making of this newsletter.
Jesper Dramsch is the creator of PythonDeadlin.es, ML.recipes, data-science-gui.de and the Latent Space Community.
I laid out my ethics including my stance on sponsorships, in case you're interested!