Roadmap to Google STEP: My Practice Plan

Preparing for Google STEP: My Practice Plan

I’m building a no-paywall, low-stress path into software by targeting the Google STEP Internship (and keeping apprenticeships open as a parallel track). This plan trades endless “grind” for systems I can repeat on busy gym days, with exact videos, links, and habits that actually stick.

My rule of thumb: Try → Hint → (short) video → close the video and re-implement from memory. Cap new problems at 25 minutes, and redo anything I needed help on at 48 hours and again at 7 days. I also practice in a doc/notes editor 1×/week (no running code) to simulate interview environments and force clear dry-runs.

My STEP Prep Roadmap

Instead of grinding forever, I’m following a 2-day kickstart and a lightweight weekly loop that strengthens resume, GitHub, and interview stories at the same time.

✅ Foundations

• Method (5–10 min skim): How to Use NeetCode Effectively — the exact learning loop I’m using.
• Problem spine: NeetCode 150 list (or Grind75/Sean Prashad; links below). I’ll pick 2 new + 1 redo per day.
• Templates I must be able to type from memory: Sliding Window, Binary Search (incl. on-answer), Monotonic Stack (next greater), BFS/DFS, and a small Memoized DP function.

🎬 Exact 2-Day Kickstart (videos + reps)

Press play, then do 1–2 reps immediately after each anchor.

Day 1 (≈3–4h)

1. Method anchor (20m): How I would learn LeetCode if I could start over (NeetCode) — locks in the try→hint→video→re-implement loop.
2. Binary Search (45–60m): Binary Search — LeetCode 704 (NeetCode) . Reps: LC 704 (classic) + a “BS on answer” problem (e.g., Koko 875).
   Hints: maintain answer ∈ [lo, hi], write a monotonic check(mid), avoid infinite loops.
3. Sliding Window (60–70m): Watch the first 2 videos in the Sliding Window playlist (NeetCode) . Reps: LC 3 “no-repeat substring” + LC 904 (or LC 209).
   Hints: expand while valid; while invalid: shrink; track counts; beware off-by-one on left.

Day 2 (≈3–4h)

1. Monotonic Stack (40–50m): Learn “next greater” using indices via NeetCode solutions playlist (see “Daily Temperatures”). Reps: LC 739 + a variant (e.g., Next Greater Element).
   Hint: you write results when you pop, not when you push.
2. Graphs — BFS (60–75m): William Fiset — BFS Shortest Path . Reps: LC 200 (islands, BFS or DFS) + LC 994 (multi-source BFS).
   Hint: mark visited on push; think in layers.
3. DP intuition (45–60m): Top 5 DP Patterns (NeetCode) . Reps: LC 70 (memo + tab) + LC 198 (or LC 322 for a stretch).
   Hint: always write: State → Transition → Base → Memo/Tab order.
4. Mock block (35–45m): Solve one medium (e.g., LC 207 or LC 424) in a notes doc: 5m restate/examples → 20–25m code (talk) → 10m dry-run → 5m Big-O/alternatives.

🔄 Weekly Loop (6 weeks, 5 days/week)

• Mon–Fri (90–120m): 2 new problems (25m cap each) + 1 redo (15–20m) + a 5m post-mortem note.
• One mock (45m): every week in a doc (no running code), to train narration and dry-runs.
• One template run: re-type a window/BS/BFS/stack/DP skeleton from memory.

🧠 Patterns To Own (STEP-level)

• Arrays/Strings, Hash Map/Set tricks, Two Pointers & Sliding Window
• Binary Search (classic + on-answer)
• Monotonic Stack (next greater / “Daily Temps” style)
• BFS/DFS (grids + topo)
• Intro DP (stairs, coin change, house robber, LIS/LCS lite)

🗒️ Hints & Insights I’m Following

• 25-minute rule: if I stall, I peek a hint, then a short video, then close it and re-implement.
• 48h & 7d redos: this is where retention happens.
• Doc-mode practice: many interviews don’t allow running code; I dry-run by hand weekly.
• Impact → how bullets: on resume/LinkedIn, start with a metric, then how I did it.
• Low-energy menu: 30–45m = 1 new (25m cap) or 1 redo + notes. No zero days.

📚 Links I Tap During the Sprint

• NeetCode — How to Use NeetCode Effectively
• NeetCode 150 list
• NeetCode — Sliding Window playlist
• NeetCode — Binary Search (LC 704)
• NeetCode — Top 5 DP Patterns (video)
• William Fiset — BFS Shortest Path
• Grind75 (customizable free plan)
• Sean Prashad — LeetCode Patterns

🧩 Minimal Project/Portfolio (Optional & Beginner-Friendly)

• learning-dsa repo: README with why I’m practicing, patterns, and a weekly goal. Add a LeetCode badge to my profile README:
  ![LeetCode Stats](https://leetcard.jacoblin.cool/YOUR_USERNAME?ext=heatmap)
• Streamlit Sprint Timer (30–45m): a one-file app that shows a 45-min timer + checkboxes: restate → plan → code → dry-run → Big-O.
• BFS Grid Visualizer (60–90m): tiny CLI or Streamlit toy; proves I’m applying DS&A, not doing ops.

🗣️ Behavioral (STAR) — My 4 Short Stories (90–120s each)

1. Tough customer → calm plan → fix → metric → one learning I reused.
2. Ambiguous ask → questions → minimal viable solution → one learning.
3. Process improvement (checklist/runbook) → fewer repeats/TTR → learning.
4. Learning fast (new tool in a week) → outcome → learning.

Why This Matters

My goal isn’t just to “pass interviews.” It’s to train like I’m already in STEP: building, collaborating, documenting, and solving in real contexts. That way when I apply, I’m not just a student grinding problems — I’m someone who ships, learns, and communicates.

Takeaway

The edge isn’t secret questions; it’s systems. This roadmap gives me a repeatable loop: daily caps, spaced redos, doc-mode practice, and targeted videos. I’ll keep both doors open (STEP and apprenticeships), tell the truth on applications, and make steady progress — even on low-energy days.