Roadmap — the whole journey

This is a living document. Reorder it whenever your curiosity pulls you somewhere — the only rule is that the daily ritual never stops.

The always-on spine

Under every season, one thing never changes: solve a problem, commit it, log it — every day. The deep track rotates; the ritual is forever. (See daily-log.md — the streak is the real metric.)

Minimum viable day = 1 problem + 1 commit. Never zero. The weekly target is the real commitment, not the daily one.

Seasons (deep track rotates; ~8–12 weeks each)

▶ Season 1 — DSA via Striver A2Z · now → ~mid-Oct 2026 · active

474 problems, 18 steps, front-loaded over 17 weeks. This is your daily ritual and your deep track right now — they're the same thing this season. → Plan · Tracker · Solutions

Season 2 — LeetCode for fun + System Design

Once Striver's done, LeetCode becomes play (contests, mediums/hards you want to do) and the deep track shifts to system design — concepts → real systems → designing and building the hard distributed primitives. → System Design roadmap

Season 3 — Mathematics for ML & Quant

The bedrock. Linear algebra, calculus/optimization, probability, statistics — rebuilt seriously, the way the later stuff demands. Start this season when the puzzles start asking for more math than you have. → Math foundations roadmap

Season 4 — Quant & Mathematical Puzzles (deep)

You'll have been dipping into the hard problem arena all along. This is where it becomes the main focus: Jane Street puzzles, the Green Book, optimal strategy, EV-under-uncertainty — as a sport. → Quant & puzzles roadmap

Season 5 — Machine Learning (core)

Implement from scratch (a neural net library, a transformer), then Kaggle, then modern frameworks. Now that the math is real, ML clicks instead of bouncing off. → ML core roadmap

Season 6+ — Go deep in one ML field

The summit. Pick one of LLMs/NLP, computer vision, RL, recsys, or ML-systems/MLOps and go genuinely deep — reproduce seminal papers, build a flagship, contribute to open source. → ML deep-dive map (+ how to choose)

Why this order

DSA first because you chose a concrete, time-boxed target and it rebuilds daily problem-solving muscle fast. System design next because it's your home turf (backend) leveled up. Then math, because it's the long-pole that unlocks both quant and ML — and you go there once the puzzles have made you want it. ML deep-dive last because it's only worth doing once the foundation can hold it.

Nothing here is locked. If month 3 you're obsessed with probability puzzles, pull Season 4 forward. The hub bends to your curiosity; it just makes sure the curiosity compounds instead of scattering.