TL;DR
- The proposed four-week plan moves from data-structure patterns to advanced algorithms and system design.
- Later weeks emphasize mock interviews, performance under time pressure, and a taper before the interview.
- Adapt the workload to your available hours and starting point, and write up what you learn from each problem.
Introduction
Most engineers do not fail FAANG loops because they lack talent. They fail because they start too late, grind the wrong topics, and peak two weeks too early or two weeks too late. Four weeks is the sweet spot for an experienced engineer who already knows the fundamentals and needs to sharpen, not relearn. This plan tells you exactly what to do on day one, on day fourteen, and on the morning you walk into the onsite.
Use it as a scaffold. Swap a topic for another if you already know it cold. Do not skip the rest days, do not front-load, and do not rewrite the plan mid-week. The timeline below assumes roughly two to three focused hours on weekdays and four to five hours on weekends. It also assumes that you already ship code in at least one language you can interview in.
Table of Contents
- Assumptions and the mental model
- Week-by-week summary
- Week 1: Pattern inventory and core DSA
- Week 2: Advanced algorithms and system design foundations
- Week 3: Mock interviews, system design depth, behavioral
- Week 4: Taper, recall, and loop day
- Daily time budget template
- Mock interview cadence
- Topic checklists by round
- Red flags that your plan is off track
- FAQ
- Conclusion
1. Assumptions and the Mental Model
The goal of the four weeks is not to learn every algorithm. It is to build three capabilities: pattern recognition under time pressure, a repeatable system design framework, and a behavioral story bank that survives a follow-up question. You are moving from breadth to depth and from practice to performance.
Think of the plan as three phases:
| Phase | Days | Mode | Target |
|---|---|---|---|
| Inventory | 1-10 | Learn and practice | Cover the pattern map |
| Pressure | 11-21 | Apply under time | Improve speed and communication |
| Peak | 22-28 | Rest and recall | Arrive sharp, not exhausted |
If you push pressure work into the peak week, you will walk into the onsite tired. If you skip the peak week, you will walk in jittery. Both are failure modes.
2. Week-by-Week Summary
| Week | Theme | Coding | System Design | Behavioral |
|---|---|---|---|---|
| 1 | Pattern inventory | Arrays, strings, hashing, two pointers, sliding window, stacks, queues | Core building blocks, reading list | Draft 6 stories with STAR |
| 2 | Advanced DSA and design foundations | Trees, graphs, DP, heaps | Scaling patterns, capacity math | Refine stories, add metrics |
| 3 | Mocks and depth | Company-tagged problems | Two full mocks per week | Record and self-review |
| 4 | Taper | Light review, recall drills | One light mock | Story fluency pass |
3. Week 1: Pattern Inventory and Core DSA
Week 1 is about mapping the terrain. You are not trying to solve hard problems yet. You are making sure every pattern exists in your short-term memory with at least one concrete example.
Daily Breakdown
| Day | Coding Topic | Problems | System Design | Behavioral |
|---|---|---|---|---|
| Mon | Arrays and hashing | 3 medium | Read: latency vs throughput | Brainstorm 10 past projects |
| Tue | Two pointers | 3 medium | Read: CAP theorem refresher | Pick top 6 stories |
| Wed | Sliding window | 3 medium | Read: load balancer types | Draft story 1 with STAR |
| Thu | Stacks and queues | 3 medium | Read: cache patterns | Draft story 2 with STAR |
| Fri | Linked lists | 3 medium | Read: SQL vs NoSQL | Draft story 3 with STAR |
| Sat | Binary search | 5 medium | 1 hour: key-value store design | Draft stories 4 and 5 |
| Sun | Review and rest | Redo 3 missed problems | Sketch two designs from memory | Draft story 6 |
Week 1 Checklist
- You can recognize the pattern from the problem statement in under 60 seconds.
- You can code a two-pointer and sliding-window template without looking it up.
- You have 6 draft stories covering conflict, failure, ambiguity, leadership, scope change, and mentorship.
- You read at least 5 system design primers.
Do not move to Week 2 if more than two of these are unchecked. Spend an extra weekend day instead.
4. Week 2: Advanced Algorithms and System Design Foundations
Week 2 introduces the algorithms that cause the most onsite failures: trees, graphs, dynamic programming, and heaps. System design moves from primitives to composition.
Daily Breakdown
| Day | Coding Topic | Problems | System Design | Behavioral |
|---|---|---|---|---|
| Mon | Trees: traversals and properties | 3 medium, 1 hard | Capacity estimation cheat sheet | Add metrics to story 1 |
| Tue | BST and balanced trees | 3 medium | Design: URL shortener | Add metrics to story 2 |
| Wed | Graphs: BFS and DFS | 3 medium, 1 hard | Design: rate limiter | Add metrics to story 3 |
| Thu | Topological sort and Union-Find | 3 medium | Design: news feed | Add metrics to story 4 |
| Fri | Dynamic programming: 1D | 3 medium | Design: notification service | Add metrics to story 5 |
| Sat | Dynamic programming: 2D | 4 medium, 1 hard | Design: chat system (2 hours) | Add metrics to story 6 |
| Sun | Heaps and priority queues | 3 medium | Review all designs, quiz yourself | Record one story out loud |
Capacity Estimation Reference
| Quantity | Approximation |
|---|---|
| Seconds in a day | 86,400, use 10^5 |
| Bytes in a KB | 1,024, use 10^3 |
| Daily active users, large app | 10^8 |
| Reads per write, social feed | 100 to 1 |
| Disk sequential read | ~100 MB/s |
| Memory read | ~100 ns |
| Round-trip across regions | 100 to 200 ms |
Memorize these numbers. They are the tokens you will manipulate during the design round.
Week 2 Checklist
- You can draw BFS, DFS, Dijkstra, and topological sort without reference.
- You can classify a DP problem into 1D, 2D, or interval in under 30 seconds.
- You can size a system to 100 million DAU with working numbers.
- Each story has two quantitative outcomes.
5. Week 3: Mock Interviews, System Design Depth, Behavioral
Week 3 is the highest-intensity week. You switch from solo practice to performance. Book at least three full mocks: one coding, one system design, one mixed. Treat every mock like the real thing: camera on, background clean, audio good.
Daily Breakdown
| Day | Coding Topic | Problems | System Design | Behavioral |
|---|---|---|---|---|
| Mon | Graphs: shortest path, MST | 3 medium | Deep dive: consistency models | Rehearse story 1 aloud |
| Tue | Backtracking | 3 medium | Deep dive: sharding strategies | Rehearse story 2 aloud |
| Wed | Mock: coding only, 60 min | Post-mortem for 30 min | Read mock feedback | Rehearse story 3 aloud |
| Thu | Tries and advanced strings | 3 medium | Design: ride sharing | Rehearse story 4 aloud |
| Fri | Bit manipulation and math | 3 medium | Design: distributed cache | Rehearse story 5 aloud |
| Sat | Mock: system design, 60 min | Review prior week's gaps | Design: metrics platform | Rehearse story 6 aloud |
| Sun | Mock: mixed loop, 90 min | Write a loop post-mortem | Review the post-mortem | Record two rapid-fire answers |
Mock Interview Rubric
| Dimension | Weight | What to watch |
|---|---|---|
| Problem decomposition | High | Do you restate the problem in your own words? |
| Communication | High | Do you narrate choices before coding? |
| Code quality | Medium | Are helpers named, is the main flow readable? |
| Testing | Medium | Do you find two non-obvious edge cases? |
| Time management | Medium | Do you finish with at least 5 minutes left? |
| Recovery | High | When stuck, do you verbalize a new angle within 2 minutes? |
Week 3 Checklist
- You completed three mocks and wrote a post-mortem for each.
- You can produce a 45-minute system design with APIs, data model, scale, and failure modes.
- Each story lands in 90 seconds or less.
- You have a short list of known weaknesses and a plan for week 4.
6. Week 4: Taper, Recall, and Loop Day
Do not learn new material this week. The taper is the hardest discipline: your instinct will be to grind. Resist it. Your nervous system needs to be fresh, and new content three days before the loop will crowd out the patterns you already know.
Daily Breakdown
| Day | Coding | System Design | Behavioral | Body |
|---|---|---|---|---|
| Mon | 2 easy, 1 medium | 30-minute design recap | Run story 1 and 4 | 30 min walk |
| Tue | 2 medium | 30-minute design recap | Run story 2 and 5 | 30 min walk |
| Wed | 1 medium, 1 revisit | Light mock, 45 min | Run story 3 and 6 | Sleep 8 hours |
| Thu | 1 easy only | Pattern doodle | Two questions aloud | Rest day |
| Fri | Rest | Rest | Rest | Sleep 8 hours |
| Sat (loop day) | Warm-up: 1 easy only | Read one design you know | Skim story bank | Hydrate, eat light |
| Sun | Debrief | Write notes from memory | No practice | Full rest |
Warm-Up Set for Loop Day
Pick one of these for the morning of. Do not pick anything that has ever frustrated you.
- Reverse a linked list
- Valid parentheses
- Binary search on a sorted array
- In-order traversal of a tree
The only purpose of the warm-up is to get into code-writing mode without bruising your confidence.
7. Daily Time Budget Template
| Block | Weekday | Weekend |
|---|---|---|
| Pattern review | 15 min | 30 min |
| Problem solving | 90 min | 150 min |
| System design reading | 30 min | 60 min |
| Behavioral work | 15 min | 30 min |
| Break and walk | 20 min | 40 min |
| Total | ~3 hours | ~5 hours |
If you cannot hit these numbers, cut the system design reading on weekdays rather than the problem solving. Pattern exposure compounds faster than reading does.
8. Mock Interview Cadence
| Week | Mocks | Focus |
|---|---|---|
| 1 | 0 | Pure study |
| 2 | 1 optional | Coding only, low stakes |
| 3 | 3 mandatory | Coding, design, mixed |
| 4 | 1 light | Confidence and pacing |
Where to find mocks: trusted peers, an ex-colleague at a target company, or a paid service. Peer mocks are fine if your peer has interviewed recently. Random pairings are hit or miss.
9. Topic Checklists by Round
Coding Round
- Arrays, hashing, two pointers, sliding window
- Stacks, queues, monotonic stacks
- Linked lists: reverse, merge, cycle
- Trees: traversals, path sums, LCA
- Graphs: BFS, DFS, shortest path, topological sort, Union-Find
- DP: 1D, 2D, knapsack, interval
- Heaps: top-K, merge-K
- Backtracking: subsets, permutations, combinations
- Tries and rolling hash
System Design Round
- Requirement clarification
- Back-of-envelope numbers
- API design
- Data model and storage choice
- Partitioning and replication
- Caching layers
- Asynchronous processing and queues
- Failure modes and graceful degradation
- Monitoring and rollout
Behavioral Round
- Conflict with a peer or manager
- A project that failed or was cancelled
- An ambiguous problem you scoped
- A decision under time pressure
- A time you disagreed with leadership
- A time you mentored or taught
10. Red Flags That Your Plan Is Off Track
| Signal | Likely Cause | Fix |
|---|---|---|
| Skipping system design daily | Overweighting coding | Cut one coding problem, read 30 min |
| Still failing easy problems in week 3 | Pattern gaps | Revisit week 1 topics, do not grind mediums |
| No mock scheduled by day 14 | Avoidance | Book three mocks today, even if imperfect |
| Stories feel scripted and flat | Over-rehearsal without metrics | Rewrite with numbers and a real conflict |
| Exhaustion in week 4 | Missed taper | Cut volume by half immediately |
11. FAQ
Q: Four weeks is not enough. Should I postpone? If you are a practicing engineer who ships code, four focused weeks beat twelve half-hearted ones. Postpone only if you cannot guarantee the daily budget.
Q: I already know DSA. Can I skip week 1? No. Week 1 is not learning, it is inventory. Compress it to four days if you must, but do not skip pattern recognition drills.
Q: How many problems in total? Around 100 to 120 problems with real write-ups. Quality beats volume. Someone solving 400 problems with no post-mortem will be outperformed by someone solving 120 with clean notes.
Q: Should I do LeetCode Premium for company-tagged problems? Yes for the last two weeks. Tagged problems bias you toward the style and patterns your target company favors.
Q: What about the recruiter call before the loop? Treat it as a fifteen-minute practice behavioral round. Have a polished one-minute self-introduction and two questions ready.
Q: I bombed a mock in week 3. Do I push the loop? Only if you bombed two in a row on the same dimension. A single bad mock inside week 3 is normal and useful.
Q: How much coffee is too much on loop day? Your normal amount, not one cup more. Caffeine spikes mimic anxiety and your body will not know the difference.
12. Conclusion
Four weeks is enough if you respect the structure. Week 1 builds coverage, week 2 builds depth, week 3 applies both under pressure, and week 4 lets your system consolidate what it has learned. The plan is deliberately boring because interview preparation is not a creative exercise: it is a loading protocol. Load, stress, taper, perform.
Print the tables. Put them on your wall. Cross off days as you complete them. By the end of week 4 you will walk into the loop with a clear pattern map, a working system design framework, and a story bank you can deliver without reading from a script. That is what a FAANG loop requires. The plan above is how engineers get there in exactly 28 days.
Frequently asked questions
Is four weeks really enough to prepare for a FAANG interview?For a practicing engineer who already ships code in an interview-friendly language, yes. Four focused weeks at two to three hours daily and four to five hours weekend beats twelve half-hearted ones. Postpone only if you cannot guarantee the daily budget or if you have not touched data structures in years.
How many mock interviews should I do in a four-week prep plan?Zero in week one, one optional coding-only mock in week two, three mandatory mocks in week three covering coding, design, and a mixed loop, and one light confidence-and-pacing mock in week four. Source mocks from trusted peers, ex-colleagues at target companies, or paid services like interviewing.io.
How many LeetCode problems should I solve in four weeks?Around 100 to 120 problems with real write-ups beats 400 with no post-mortem. Quality of reflection matters more than volume. Use LeetCode Premium for company-tagged problems in the last two weeks since tagged problems bias you toward your target company's style.
What should I do during the taper week before FAANG interviews?No new material. Cut volume by half. Do two easy and one medium problem on weekday mornings, run thirty-minute design recaps, rehearse two stories per day, and sleep eight hours the two nights before. Loop-day warm-up is one easy problem you have never been frustrated by, like reverse a linked list or valid parentheses.
Should I skip week one if I already know data structures cold?No. Week one is inventory, not learning. The goal is making sure every pattern lives in your short-term memory with at least one concrete example so you recognize it in under sixty seconds during a real round. Compress the week to four days if you must, but do not skip pattern recognition drills.