Why I built AlgoPath

I could write code. I had shipped things real people used. I still sat in front of an interview problem. No idea what to do. AlgoPath exists for that gap.

What was actually wrong?

I did what everyone recommends. Pick a problem, get stuck, read the solution, feel fine. Then get stuck again.

Around problem 60 I noticed something. I was not getting better at solving. I was getting better at reading solutions. Only one of those shows up in an interview.

Why that habit forms is in why people hate algorithms.

When did it become obvious?

A problem asked for the longest stretch under some limit. I stared at it for 20 minutes. The answer was a sliding window. I had solved four of those that same week.

I knew how to write one. Spotting one was the part I could not do. No tutorial had ever covered that step. They all started after it.

What is the step nobody teaches?

Solving an algorithm problem has three parts. Work out what kind of problem it is. Recall how that kind is solved. Write the code.

Courses teach parts two and three. Problem sites give you three. Part one decides interviews, and it is left to accident.

Part one arrives after a few hundred problems. Most people quit first.

Training that step on purpose is pattern recognition.

What did I build?

150 steps in order. Each takes one idea, explains it, and makes you write it. Then it lands in a real system. A rate limiter, an LRU cache, an autocomplete.

Alongside them: 52 patterns, each with the signals that point at it. And 194 problems with the label removed. You name the pattern before writing a line.

The unlabeled problems build a map of what you confuse with what.

How does one step work?

Learn the signals before the code, on purpose. Write the core yourself in the browser, in JavaScript, Python or PHP. Then wire it into the real case.

A step is done when you can explain it with no notes. And rewrite the core from a blank file. That is the testing effect doing the checking.

The step comes back days later, then weeks, then a month. Plain spaced repetition.

Why is every lesson written three times?

You pick the depth, step by step. Simple assumes nothing. Standard is how a working developer would say it. Academic gives invariants and precise complexity.

Write only for beginners and everyone else leaves on page one.

Who is it for?

People who can write a loop and a function. And stall the moment a problem stops announcing what it wants.

Career switchers and self-taught developers. People with a few years of work and no formal study. I am in that last group.

This track is not for competitive programming. It will not get you through an interview next week.

Why am I the one writing it?

I am Elman Huseynov. Ten years as a software engineer. Zalando in Berlin now, Shopify and Scoutbee before that. More at helloelman.com.

None of that stopped me freezing. That is the part worth saying out loud.

I never studied computer science. Everything I know came from shipping, which leaves holes. This was one of them.

I built the course while learning it. Every place the explanation slows down is a place I got stuck.

Where do you start?

The first 33 steps are free in full. Big O, hash tables, two pointers, sliding windows, binary search.

Start at step 1 if those words are new. Take the placement test if you already write code. It takes seven minutes and no account.

The whole map is on the roadmap. AlgoPath is at algopath.pro.