3,178 karma · joined September 2, 2007
DSR is the market leader in EU for machine learning training. DSR helps coders or people with significant quantitative training (e.g. science, engineering, or math graduates) ramp-up rapidly for a data science career - arguably the fastest-growing, highest-demand profession. Our participants have an average of 5 years of industry experience.
DSR is a 3-month, in-person, rigorous, and full-time/intensive course in the startup-capital of Europe: Berlin. You’ll learn software engineering, data science, business analysis and communication faster and more deeply with mentors doing code reviews and pair-programming - all on real-world data and problems. You will develop a portfolio project, demonstrating you can own a business problem, solve it, and communicate why your results are definitive.
You bring your training, tuition, and drive to master our curriculum alongside our world-class mentors and partners. Towards the end we provide a networking event with top-tier technology companies, where you will show off your new skills and portfolio project, likely leaving with a career-changing job.
From my limited experience, Canada is somewhat a middle point culturally between US and EU. This aspect (debt) seems to be far closer to the US, at least when compared to northern EU where average debt per family is virtually zero.
Data Science Retreat (DSR) is the market leader in EU for machine learning training. We are expanding to North America, we started in SF but are moving to Montreal. DSR helps coders or people with significant quantitative training (e.g. science, engineering, or math graduates) ramp-up rapidly for a data science career - arguably the fastest-growing, highest-demand profession. Our participants have an average of 5 years of industry experience.
We are launching a new product that is a big departure from our 3-month in-person training:
- For a monthly fee (3 different price points), our customers receive a topic that our teachers prepare in depth. For example, an exciting new technique that just came out in a paper and that is really applicable in the industry today
- There's code that implements the technique, but the most important part is that we provide a detailed explanation and references to every bit of knowledge you need to understand the topic. With exercises that are self-correcting (you write code, press a button, get red/green)
- This business model works well B2C, but it's something easy to consume for big corporates that want to train their people on-the-job (the current 3-month retreat doesn't address this need well). It's all very 'hands on' and more advanced than the usual MOOCs. We suspect MOOCs are not creating data scientists, at least not when done on-the-job. If we can demonstrate results, this should be an easy sell.
As Director of this product, these could be your day to day activities:
- Think up methods to increase our pipeline, together with marketing - Pay attention to KPIs for recurrent business, for example, churn - Decide which proportion of revenue goes to service the base, and which to invest in R&D - Look at customer lifetime value (you don't need to calculate it, we have a resident expert on the team) - Experiment with variations of product and pricing - Do outbound sales to convince companies to try our service
Side gig for someone who loves teaching and mentoring. Can be on a weekend day, make sure your contract with your fulltime employer allows that. If you don't know deep learning to a level you can teach it but know classical ML, you are good. You are a specialist in one ML domain. You're a good communicator, you're self-sufficient & you're passionate about the work you do.
==Mission==
Take 'Bedroom deep learning' to the next level. Every participant comes out of the door with a portfolio project that has social impact, ideally at scale.
At DLR, we believe (in spite of the hype) deep learning is actually underappreciated. It has more potential to generate an impact than most other technologies. Big companies (Google, Amazon, Facebook, Netflix, Waymo, Uber, Apple, etc.) are ‘killing it’ using deep learning in their products. You hear you need lots of data and computation to build anything remotely useful. This is just not true. You don’t need to be a corporation to have ridiculous amounts of effect with deep learning!
The more we talk to companies interviewing today, the more apparent it is: A portfolio project is decisive when making hiring judgments. Jeremy Howard recommends it. Andrew Ng recommends it. Why? It’s far better at discriminating talent than any other proxy (CVs don’t work; pedigree doesn’t work; puzzles don’t work).
You can read more about our method here: https://deeplearningretreat.com/method-and-manifesto/ See also the kind of projects we build: a combination of cheap hardware and deep learning that produces serious social impact.
==Company Values==
https://deeplearningretreat.com/work-at-dlr/ We follow the 'Teal' model (Book: reinventing organizations). The company has a strong purpose and everyone is autonomous and empowered.
Other instructors tell us that they learn a lot by teaching here. You can join anyone else's session anytime.
We are customer funded.
==Interview==
Video call (1 hr) >> Test (3 hrs) >> teaching to DLR participants (half-to-full day)
Side gig for someone who loves teaching and mentoring. Can be on a weekend day, make sure your contract with your fulltime employer allows that. If you don't know deep learning to a level you can teach it but know classical ML, you are good. You are a specialist in one ML domain. You're a good communicator, you're self-sufficient & you're passionate about the work you do.
==Mission==
Take 'Bedroom deep learning' to the next level. Every participant comes out of the door with a portfolio project that has social impact, ideally at scale.
At DLR, we believe (in spite of the hype) deep learning is actually underappreciated. It has more potential to generate an impact than most other technologies. Big companies (Google, Amazon, Facebook, Netflix, Waymo, Uber, Apple, etc.) are ‘killing it’ using deep learning in their products. You hear you need lots of data and computation to build anything remotely useful. This is just not true. You don’t need to be a corporation to have ridiculous amounts of effect with deep learning!
The more we talk to companies interviewing today, the more apparent it is: A portfolio project is decisive when making hiring judgments. Jeremy Howard recommends it. Andrew Ng recommends it. Why? It’s far better at discriminating talent than any other proxy (CVs don’t work; pedigree doesn’t work; puzzles don’t work).
You can read more about our method here: https://deeplearningretreat.com/method-and-manifesto/ See also the kind of projects we build: a combination of cheap hardware and deep learning that produces serious social impact.
==Company Values==
https://deeplearningretreat.com/work-at-dlr/ We follow the 'Teal' model (Book: reinventing organizations). The company has a strong purpose and everyone is autonomous and empowered.
Other instructors tell us that they learn a lot by teaching here. You can join anyone else's session anytime.
We are customer funded.
==Interview==
Phone call / coffee (1 hr) >> In-person (3 hrs) >> Onsite w/team, teaching to DLR participants (half-to-full day)
Deep learning Retreat (DLR) is the first deep learning school in the world. We train people who want to either:
- Have a job in a tech company
- Have tons of impact by going alone on their own startup or doing ‘good’
We make sure we fulfill those two customer needs by giving them the best environment to produce an impressive portfolio project.
The more we talk to companies interviewing today, the more apparent it is: A portfolio project is decisive when making hiring judgments. Jeremy Howard recommends it. Andrew Ng recommends it. Why? It’s far better at discriminating talent than any other proxy (CVs don’t work; pedigree doesn’t work; puzzles don’t work). DLR focus is to help you write the best portfolio project you could; in that sense, it’s like a writer’s retreat, where one goes to finish your masterpiece novel. We use online videos to cover the theory, but focus on project-based, peer-learning methods that are completely hands on. When you come out of the door, you should be holding a project that proves you can write damn good code and makes you unstoppable in interviews… or an extremely desirable partner for social impact projects or startups.
I think you took the 'webapp' UX model a bit too far (where everything must be click, animation, multiple choice. Not even right click). Few modern web apps do these things well. One is checkvist.
Your target user likes the keyboard (coders, writers). Listen to them (us)
- Hmm, no shortcuts? or keyboard support in general. Press arrow down to highlight a search result fails
- bounding box for a note not needed (see onenote)
- Search the content of a note not possible?
- Too much chrome while note writing?
These are simple things to fix. Great project!
One unorthodox way to check the effect of a diet: "Did you feel any better?"
I'm currently vegan. I don't fell better. The only time in my life when the diet seemed to have an effect I could notice in mood and energy levels was when I was 'raw'.
I don't use N26 because they don't have business accounts, but I would if they did.
If we keep consuming news like this, and do nothing, it's going to scalate massively. Same way as when Snowden told people they were spyed on and they collectively shrugged and continued with their lives as if nothing had happened.
We, people in tech, have a massive moral burden to educate 'normals' on the meaning of news like this!
For an orthogonal set of arguments that benefit the US, check 'the accidental superpower' (Geopolitics and demography make the EU and CN not likely to outdo the US all things considered; unless the US shoots itself in the foot with say a civil war)
Would you take a machine learning course from an unknown, likely outdated professor when you could listen to Andrew Ng for free?
What the internet has brought is a 'winner takes all' for education.
Andrew has the reputation to do this and not be hated. But you know who did exactly this? Rocket internet, the clone-war masters.