100 Days of Machine Learning
I recently came across a Medium article on completing 100 days of Machine Learning, and it struck a chord with me. Not because it promised mastery in a fixed timeframe, but because it emphasized something far more important: discipline. That’s what I want to build this year (2026)
Over the next 100 days, I’m committing to showing up daily to learn and build in the ML/AI space. My focus will be on three pillars:
- Foundations & concepts - revisiting core ML ideas and understanding why things work
- Research exposure - reading papers, summaries, and implementations to stay grounded in real-world progress
- Hands-on projects - practical implementations, even if they’re tutorial-based
I’m not chasing novelty. I’m chasing depth and consistency.
One personal rule I’m setting for myself: every project will include a “community bonus” section. This is where I’ll reflect on how the idea, technique, or system could be applied to problems that matter to my community - whether that’s education, accessibility, public services, or local decision-making. Even hypothetical links count, as long as the thinking is intentional.
My long-term goal is to become a stronger ML engineer - someone who can reason deeply about models, data, and trade-offs - and eventually work on problems with direct, positive impact. This 100-day challenge is a small but deliberate step in that direction.
I don’t expect perfection. I do expect progress.
Day 1 starts now.
Update (10th Jan 2026)
I came to realise to learn in depth, I might need to write in cycles - each cycle is a deeper dive into the ML project, research etc.
It might take weeks to complete 1 cycle but each time there will be a proper repository, artifact or report being written as opposed to daily posts which might be more superficial. So do look out for more of these in the upcoming weeks!