Learning

Advices, Best Practices, Principles

Learn in Public

• Continuously seek feedback and be open to learning
• Demonstrate your value through action and collaboration.

Advice: Build things from the ground up

Constanstly use knowledge sharing Get Your Hands Dirty Do many projects, start small. Incremental improvements.

Learning examples as a well-focused projects with a paper and documentation.

Engage more deeply with topics by experimenting more; start by crafting simple concepts and connecting them effectively.

• you need to constantly apply learning principles (and apply)
• always remember your ‘why?’

“What I cannot build, I do not understand” Richard Feynman

Consistency is key: dedicate at least one hour or more each day to coding.

grow habit of creating Learn in public

Practicing is really important.

Practicing

Practicing generally refers to the act of repeatedly doing something in order to develop a skill, improve performance, or gain mastery. It involves consistent effort and dedication over time. For example, practicing a musical instrument means regularly playing scales, exercises, or pieces to become more proficient. Practicing a sport involves going through drills, techniques, and gameplay strategies to enhance agility or teamwork. Essentially, to practice is to engage in an activity repeatedly and deliberately with the goal of growing one’s expertise or ability.

practicing vs application vs usage

best practices and advices:

connection with improvement

practicing objects…

top down vs botton-up approach

Constructivist Learning

• Define main principles and use them

https://en.wikipedia.org/wiki/Experiential_learning

to_sort

https://en.wikipedia.org/wiki/Learning_styles#Criticism

https://en.wikipedia.org/wiki/Kolb%27s_experiential_learning https://en.wikipedia.org/wiki/Learning_cycle

https://en.wikipedia.org/wiki/Corrective_feedback

https://www.youtube.com/watch?v=ddq8JIMhz7c

High-Impact Practices George D. Kuh identified High-Impact practices (HIPs) as " a Specific set of practices that tended to lead to meaningful experiences for students."

learning / understanding has different meanings for different things. So, ok. What else? -> Collect and practice LN for LB, methods, processes. Iterate PDCA/PDSA.

Memorise some common tradeoffs (you’ll see some patterns once you learn about 7-8 designs)

how to make a notes? How to make them useful? what do you need to note and how?

kind of “systemic” approach - in the end you should give working prototypes with all other things (documentation, reasoning)

how would you learn some thing: sources (articles, papers) + notes + practical usage reflection (in the cases list - you need to create a list of them)

Life cycle of a thing to use a Feature (thing, tool, method, practice etc. ): learn -> modify -> practice -> analyse -> teach and get feedback -> improve

Social Bot: learn > publish > modify->use->analyse&publish-> discuss & get feedback

how to effectively get feedback?

• use specific examples
• teach others what you have learned
• socializing
• writing posts about this
• repeating, practicing, combining
• decomposition & small task progress

List:

• communicate, teach, collaborate
• solve problems (amazon thing)
• feedback
• practical and well-focused projects (one main target)
• problem-first approach for some of the topics

glossary / (other word)

terms and things: roadmap, plan, steps, guides, cirriculum,

methodologies

reflexive_learning

and connections with Reflective_practice

IJSDL10.1 self-directed learning

knowledge_sharing

Offer to host workshops or lunch-and-learn sessions on topics bridging DevOps, architecture, and ML. Share insights from your learning journey or projects you’ve undertaken.

launch and learn sessions as an example of knowledge sharing. how they’re organized review Launch And Learn https://trueaccord.atlassian.net/wiki/spaces/ENG/pages/1442185233/2024+Lunch+and+Learns

discussion platforms; discussions;

https://en.wikipedia.org/wiki/Double-loop_learning also triple-loop learning

Lessons Learned

https://en.wikipedia.org/wiki/Lessons_learned book: The Lessons Learned Handbook (file) #todo worth reading file: eput lessons team (some interesting things)

Lessons learned (American English) or lessons learnt (British English) are experiences distilled from past activities that should be actively taken into account in future actions and behaviors.

Communication and Collaboration

discuss problems in https://www.reddit.com/r/MLQuestions/

How to land a job – go to ML panel events. Talk to the panelists. Ahead of time, do some research on the panelists so that when you chat with them, you can impress them and feed into their ego that you know what they’re working on. If you have also done some side projects, be sure to mention it. They’ll likely ask for a resume and see if you’re a good fit. Landing a job requires both knowledge of the ML and being a hustler and networking with folks. Most of people I know who landed an ML role after working in traditional software engineering simply networked and knew the right people. You can network from anywhere in the world as long as you have an internet connection.

? collaborative learning <- collaborative problem solving, etc.. collaborative? collective

Offer to host workshops or lunch-and-learn sessions on topics bridging DevOps, architecture, and ML. Share insights from your learning journey or projects you’ve undertaken.

Definitions

Learning is the process of acquiring new understanding, knowledge, behaviors, skills, values, attitudes, and preferences.

Here are various definitions of learning from multiple perspectives, each emphasizing a unique aspect:

1. Cognitive Definition: Learning is the process by which individuals encode, store, and retrieve information, using it to make sense of the world and solve problems. It focuses on internal mental processes and knowledge acquisition.

2. Constructivist Definition: Learning is an active, constructive process where individuals create meaning from experiences. This view suggests that learners build knowledge by integrating new experiences with prior knowledge through exploration and reflection.

3. Connectivist Definition: Learning is the ability to make connections within networks of information sources, often facilitated by technology. This view, designed for the digital age, focuses on networking, adaptability, and finding relevant information in complex systems.

4. Biological Definition: Learning is the formation and strengthening of neural connections in response to new stimuli or experiences. This perspective considers learning as a physical process within the brain, driven by neuroplasticity.

5. Educational Perspective: Learning is the acquisition and application of knowledge, skills, values, and attitudes essential for personal, social, and professional development. It highlights structured instruction and lifelong growth.

6. Experiential Learning Definition: Learning is the process of gaining knowledge or skills through direct experience and reflection, involving active engagement with real-world challenges (as per David Kolb’s Experiential Learning Theory).

7. Informal Learning Definition: Learning is a spontaneous, unstructured process through which individuals acquire knowledge from everyday interactions and experiences, outside of formal education or training.

Each definition underscores different elements, like behavior change, cognitive processing, social influence, network connections, brain adaptation, and real-world applications, showing how learning is a multifaceted and adaptive process.

learn_in_public

https://www.swyx.io/learn-in-public

v1-principles-learn-in-public.pdf

There’s a lot to learn, and you will never be done learning, especially with AI, when new revolutionary papers and ideas are released weekly.

The biggest mistake you can make is to learn in private. You don’t create any opportunities for yourself if you do that. You don’t have anything to show for it besides being able to say you completed something. What matters more is what you made of the information, how you turned it into knowledge to be shared with the public, and what novel ideas and solutions came from that information.

That means having a habit of creating.

To escape tutorial hell and really learn, you have to get hands-on, write algorithms from scratch, implement papers, and do fun side projects using AI to solve problems.

• use proper theory , use a theory, experimentation, modeling, simulation..

Do many projects, start small.

This can mean:

• writing blogs and tutorials
• join hackathons and collaborate with others
• ask and answer questions in Discord communities
• work on side projects you’re passionate about
• tweeting about something interesting you discovered new

And speaking about Twitter,

ml#

Organization Studying online courses for hours per day can be hard, it’s very active engaged learning. I’ve found 6 hours on days off and 2-4 hours on work days is a nice middle ground. I usually read 2 hours, work on math for 2 hours, work on ML courses for 2 hours. I’ve had a couple of nice work related data science projects that I fully commit time to when they come up. I always apply methods to my own datasets and build my own implementations alongside the coursework.

I literally learned most of the bases from chatgpt. Then practiced on kaggle and trying to replicate papers result or models

If you want to improve your knowledge in a certain area, I’d actually recommend making a project and looking up whatever issues you face and whatever you need to learn for it. But for starting out, it is horrible. I’d definitely recommend learning from traditional courses first, starting with linear algebra and statistics, then move on to the basics of machine learning.

learning this curriculum may be extremely challenging to take on alone. It is highly recommended to find a more experienced mentor, or at the very least a study partner. There are suggestions for more accessible alternatives, which have the same prerequisites, below.

https://medium.com/bitgrit-data-science-publication/a-roadmap-to-learn-ai-in-2024-cc30c6aa6e16 interesting one with a lot of advices

https://towardsdatascience.com/how-id-learn-machine-learning-if-i-could-start-over-c68d697e6a8a https://medium.com/geekculture/how-to-become-very-good-at-machine-learning-a78e3d93fe96 https://medium.com/@amirziai/a-semi-random-collection-of-advice-that-ive-found-very-useful-as-an-ml-engineer-d7ae59209a76

how to learn https://www.reddit.com/r/learnmachinelearning/comments/1fre362/a_note_to_my_six_month_younger_self/

topic: behavioral questions. learning is also here

Here is another type of learning: learn from mistakes learn from incidents to prevent future problems and improve (what?)

Understanding

Understanding

Understanding is a relation. We can model it.

Common Learning

Problem-Based Approach

Problem-based learning #learning/principle

problem-first approach https://mkremins.github.io/blog/doors-headaches-intellectual-need/

firstly you need to understand a problem, a limitations of currently using tools and then, analyzing several possible solution options you have to construct your own #learning

https://en.wikipedia.org/wiki/Problem-based_learning https://en.wikipedia.org/wiki/How_to_Solve_It

Project-Based Approach

parallels with problem-based, formulate differences

https://en.wikipedia.org/wiki/Project-based_learning

advice from ML: If you want to improve your knowledge in a certain area, I’d actually recommend making a project and looking up whatever issues you face and whatever you need to learn for it. But for starting out, it is horrible. I’d definitely recommend learning from traditional courses first, starting with linear algebra and statistics, then move on to the basics of machine learning.

I literally learned most of the bases from chatgpt. Then practiced on kaggle and trying to replicate papers result or models

https://www.reddit.com/r/learnmachinelearning/comments/1d1u2aq/i_started_my_ml_journey_in_2015_and_changed_from/ “If you already know the topics that Andrej Karpathy covers in zero to hero, and you have applied them at work, you are already ahead of lots of FAANG ML engineers. If you’re feeling bored or hitting a ceiling, it’s time to start thinking about what problems you’d like to tackle at work or in the world and how you’d solve them with ML. You can then start to tweak your models to solve these problems. Real world ML requires a lot of data acquisition and sanitizing before you even get to the modeling step. Once you’re ready to develop a model, there are so many directions you can take it.” <- here modifying and applying cycle (not full one)

Personally I find that I need a substantial project to implement in a new interesting language. You can read a book at become familiar with syntax, and be exposed to interesting features, but a nothing gets you thinking in a new language like a project.

You can do 1-2 YouTube tutorials just to get a feel for it if that’s helpful to you, but you won’t learn nearly as much this way. It’s important to try to build something yourself as soon as possible. You’ll feel overwhelmed by not knowing what to do, but just read the documentation, google what you don’t know, and keep making progress forwards even if it feels slow

Creation Lens

As a backend developer, you can apply creation as a method to deepen your understanding, improve your skills, and become more effective in your role. Here’s how you can approach it in a practical and structured way.

Understanding isn’t just theoretical; it must translate into the ability to create or replicate the concept.

Feynman’s Approach in Action

Creation is a way to “test” your understanding against reality. Feynman believed that if you could explain (or in this case, build) something from the ground up, then you truly understood it. If you struggled, it was a signal to revisit and refine your knowledge.

Learning Through Construction

The process of building forces you to confront gaps in knowledge and strengthens your understanding.

Software Development

• If you can’t write code to implement a feature or system, it may indicate you don’t fully understand the underlying principles.
• Backend developers, for instance, need to grasp algorithms, databases, or APIs well enough to construct and optimize them.

Problem-Solving

The ability to “rebuild” a solution from scratch is often the ultimate test of comprehension.

To Understand Complex Systems

• Building something forces you to break down complexity into manageable, understandable parts.
• It helps uncover hidden interdependencies, nuances, and challenges that theory alone cannot reveal.
• Example: Writing a backend API helps you understand how data flows through systems and why certain architectural choices are made.

To Identify Knowledge Gaps

• The process of creation highlights areas where your knowledge is incomplete or assumptions fail.
• It encourages asking critical questions and seeking out resources to fill those gaps.
• Example: While implementing a new feature, you might realize you lack understanding of certain database query optimizations.

? Internalize learning Learning a programming language becomes solidified when you build a project rather than just reading documentation.

To Communicate Understanding

• The ability to build something demonstrates understanding in a way words cannot.
• It’s a practical way to show others how something works, making knowledge transferable and shareable.
• Example: Creating a proof of concept or demo showcases your understanding of a technology to stakeholders.

To Build Confidence#

• Seeing your ideas take shape boosts confidence and motivates further exploration and learning.
• Example: Successfully deploying a project reinforces your belief in your capabilities and encourages tackling larger challenges.

Also - to push boundaries and to innovate

• Creation fosters innovation by combining existing knowledge in new ways.
• The act of building inspires creativity and often leads to unexpected insights.
• Example: Building a microservice architecture for a specific use case might inspire a novel way to improve scalability for other systems.

Experimentation#

Experimentation: Testing Ideas and Learning Through Failure(?)

Edge Cases Simulate extreme scenarios like high traffic, large datasets, or service failures. Experiment with throttling, retries, and circuit breaker patterns.

Mindset for Experimentation#

• Fail Fast, Learn Faster: Expect failure and treat it as a learning opportunity.
• Document Observations: Keep a record of what worked and what didn’t to improve iteratively.
• Stay Curious: Ask “What if?” questions to push boundaries

Applications

[speaking]

System-design

Go Learning

be a mentor on https://exercism.org/

Here we destructuring (other word) the topic of learn to pieces: the pieces of a language: syntax, structures, standard libs, collections, ecosystem

common engineering practices patterns

tasks, exercises, katas, lessons, common methodology: https://www.reddit.com/r/learnprogramming/comments/uh96pf/fastest_way_to_learn_a_new_programming_language/ https://simpleprogrammer.com/learn-new-programming-language/