Guide: Managing Energy During Interview Loops

Who this is for

An engineer preparing for a 4–6 hour interview loop in the next 2–4 weeks who has noticed that “study more” stopped working a few loops ago. You can solve the problems in isolation. You’ve read the walkthroughs. The thing that breaks is round 3 or 4, when your working memory has thinned out and the candidate who showed up at 9am isn’t the candidate sitting in front of the system design whiteboard at 1pm.

This is especially the guide for ADHD and autistic engineers — for whom cognitive energy is a more visible, more depletable resource — but the model applies to any engineer for whom a loop costs more than the hours on the calendar.

It’s a companion to Surviving a Full Interview Loop with ADHD, which covers prep systems and in-round mechanics. This piece is about the resource the systems run on.

The core insight

Most interview prep advice treats time as the constraint: study more hours, do more problems, sleep more. Useful, partially true, and missing the actual bottleneck.

The bottleneck for senior loops is cognitive energy, not time. It behaves more like glucose than money: you can’t bank it across days, you spend it whether you intend to or not, and the curve through a loop is steep enough that round 4 is a different person from round 1.

The reframe: an interview loop is a 14-day energy budget that ends on loop day. Not 14 days of study with one big day at the end. Where that energy goes in days T-14 through T-1 determines whether loop day has a working candidate at the table.

Three corollaries fall out of this:

1. The fight isn’t to add prep — it’s to subtract drain. A 30-minute meeting that depletes you is more expensive than a missed prep session. Most candidates ignore this and over-spend on prep.
2. The taper matters more than the peak. Athletes know this. Engineers don’t. The 72 hours before loop day are a recovery window, not a cramming window.
3. Loop day’s job is preservation, not optimization. You will not get smarter on loop day. You will only get more or less depleted. The goal is to spend down evenly, not peak in round 1.

Standard interview-prep advice (“just sleep well, eat breakfast”) fails because it treats loop day as the only relevant day and energy as a binary: had it / didn’t. The actual problem is shape — where the curve sags.

The tactics

1. Map your real energy curve before you plan anything

Most prep schedules assume a flat day: 8 hours of equivalent cognitive capacity, evenly distributed. Yours isn’t flat. Mine isn’t either. Before you build a prep schedule, build a one-page sketch of your actual daily curve — the times you can do hard work and the times you can’t.

A useful exercise: for the next three workdays, log your energy on a 1–5 scale every two hours. You’ll see a shape — almost certainly a peak in some 2–3 hour window, a dip after lunch, a smaller peak in late afternoon for some, evening for others. For ADHD engineers on stimulants, the shape is sharper and more predictable; for autistic engineers, it’s heavily affected by the prior day’s social load.

The output you need: one sentence each for peak window, recovery window, do-not-attempt window. Example: “Peak 9–11:30am. Recovery 1–3pm. Do-not-attempt after 7pm.” Then check loop day’s schedule against this. If the loop is 10am–3pm and your do-not-attempt window starts at 2pm, you have a structural mismatch and should ask recruiting to adjust scheduling. Most won’t fight this.

2. Treat the 14 days before the loop as an energy budget, not a study schedule

The standard prep schedule looks like: Week 1 study, Week 2 study harder, Week of loop study hardest, Day before loop “rest”. This inverts what should happen.

A more honest 14-day shape:

Two-thirds of candidates I’ve talked to violate the taper. They study hard until 11pm the night before. They do mock interviews on T-1. They open a new system design textbook at T-2 in panic. Each of these moves trades 3% confidence for 15% loop-day depletion. The trade is bad and it compounds.

3. In the prep weeks, subtract drain before adding study

The standard advice is “study more.” The advice that actually moves the needle is “spend the budget you have.” Most engineers entering a loop are running an energy deficit before prep starts — too many meetings, on-call rotations, an exhausting product release, a 3-year-old at home. Adding study on top of that deficit produces diminishing returns within days.

Aggressively cut for the 14-day window:

• Move on-call. Trade with a teammate. Most teams will swap if asked; almost no one asks. The cost is one favor; the benefit is about 4 hours per night of recoverable sleep across two weeks.
• Decline non-load-bearing meetings. “I’m in interview prep, can we async this?” is a complete sentence. Most managers understand. The ones who don’t are signaling something useful.
• Push off social commitments that require performance. Dinner with close friends — fine, even restorative. Work happy hour, a big family event, a wedding — those are cognitive energy spends disguised as recovery.
• Pause one ongoing exhausting habit. Whatever it is — a side project, a podcast you’re producing, a workout block that’s been depleting you. One thing. Pick the one with the worst ratio of obligation to benefit.

The savings here typically dwarf the marginal benefit of two extra hours of LeetCode per week. Most engineers can find 8–15 hours of recoverable budget in the 14-day window without touching prep.

4. Run the 3-day taper deliberately

The 72 hours before loop day are not for cramming. They are for recovering enough executive function to make round 4 viable. The taper is the single highest-leverage move and the most commonly botched one.

What to do in the taper:

• Light review only. Re-read your behavioral stories out loud once. Re-skim two system design walkthroughs you already know cold. Solve one warm-up coding problem per day, not three. The goal is staying warm, not adding capacity.
• Sleep on a fixed schedule. Same bedtime, same wake time, all three nights. Sleep debt compounds non-linearly across multiple bad nights. One bad night before the loop is recoverable; three is not.
• Reduce input volume. Less news, less social media, less long reading. Working memory recovers faster when it’s not being fed novelty. This is true for everyone; for ADHD and autistic brains it’s load-bearing.
• Do something physically restorative on T-1. A walk, a swim, an easy bike ride. Not a workout that depletes you, not a rest day that lets you spiral into anxiety. Something parasympathetic for 30–45 minutes.

What not to do:

• Schedule a final mock interview on T-2 or T-1. The information it generates is unactionable; the depletion it generates is real.
• “Just look over” a topic you don’t know well. If you don’t know it cold by T-3, you don’t know it for the loop. Trying to learn it now eats the taper.
• Drink alcohol “to relax” the night before. Sleep architecture hits next-day executive function harder than the subjective relaxation justifies.

5. Treat loop day as a spend, not a peak

The model most candidates have for loop day is peak performance — walk in fully charged, deliver maximum capability for 5 hours, collapse. The energy reality is closer to spend down evenly — walk in with 100 units of capacity, lose 20 per hour, finish around 10. Round 1 doesn’t need 100. Round 5 needs more than 0.

Practical implications:

• Don’t peak in round 1. Some candidates over-perform in round 1 by 30%, burn through their working memory budget, and arrive at round 3 already at 50%. Round 1 is a calibration round — competent and steady, not a showcase. The interviewer is not comparing you against you-at-100%; they’re comparing you against the bar.
• Take the breaks. A 10-minute break is not “wasted time.” It’s the recovery that keeps round 4 viable. Put your phone away, drink water, do nothing. Doom-scrolling burns more executive function than the round before it.
• Eat once, between rounds 2 and 3. Protein-and-fat heavy. Not a sandwich (carb spike → 30 min crash → round 3 collapse). Not a coffee (the caffeine half-life puts the crash in round 4 or 5). A handful of nuts and a banana is structurally better than a pastry, and counterintuitively better than a “real meal.”
• Reserve nothing for after. Whatever’s left at the end of round 5, spend it. This is the inverse of pacing in a workout. You can crash hard at 4pm. You can’t redo round 5.

6. Build micro-recoveries that actually work between rounds

The 5–15 minute gaps between rounds are the single biggest energy lever inside loop day. Most candidates fill them with phone-checking, mental replay, or rumination — all of which deplete rather than recover. The recovery has to be deliberate.

What works for most engineers in a 10-minute gap:

What doesn’t work:

• Replaying the round that just ended. You can’t change it; the cognitive cost of replaying is real and the benefit is zero.
• Checking Slack or email. Context-switching is precisely the expensive operation you’re trying to avoid.
• A coffee mid-loop. The crash window aligns with round 4 or 5.
• Cramming for the next round. If you needed it, you’re already past the point where it helps.

The ADHD-specific note: the dopamine pull toward your phone in a break is strong and the cost is non-obvious until round 4. If you have to physically put the phone in your bag in another room to make this work, do that.

7. Decide accommodations on cost, not stigma

Many companies offer interview accommodations for documented ADHD, autism, anxiety disorders, and chronic conditions: extended time, extra breaks, split-day scheduling, written instead of verbal problem statements. The decision of whether to ask is often made on emotion (stigma fear) rather than the actual cost-benefit.

The honest framing: the cost of asking is one form, low. The benefit varies by accommodation and by you. Make the call on benefit:

• Ask if it would materially help. If you’re a parent of small kids and a 5-hour loop without a real lunch break is going to destroy round 4, ask for split-day scheduling. The benefit is large, the cost is one email.
• Ask if you have a clear documented need. Stimulant timing not aligning with the loop window. Sensory issues with the proposed location. A condition that affects sustained attention. Document, ask, move on.
• Don’t ask for a small benefit at a large cost. If the accommodation is marginal and asking would mean disclosing something you’d rather not, the cost-benefit doesn’t favor asking. Plenty of people do well without accommodations they’d technically qualify for.

Two things worth knowing:

1. The companies most likely to handle accommodations professionally are the ones with the most structured loops in the first place. The same processes that produce calibrated rubrics also produce functional accommodations workflows. This means the cost of asking at top-tier companies is structurally lower than the ambient stigma fear suggests.
2. You don’t owe anyone your diagnosis. Most accommodations require only a doctor’s note describing the accommodation needed, not the underlying condition. Recruiters typically don’t see the diagnosis. Ask the recruiter directly what they need; the answer is usually less than people expect.

If you’d actively prefer not to disclose, the structural changes in this guide (taper, energy budget, micro-recoveries, schedule against your real curve) can recover most of the same benefit. They are not equivalent to a real accommodation, but they’re not nothing.

8. Plan the post-loop crash

The 24–72 hours after a loop are the part most candidates skip in prep planning. They show up unprepared and the consequences are real: poor recovery means the next stage of life (the next loop, the next sprint at work, parenting, anything) starts from a deficit.

Plan three things explicitly before loop day:

• Block the evening of loop day. No work. No more studying. Nothing requiring executive function. Pre-decide a low-effort evening activity (TV, a meal you’re not cooking, a walk) so you don’t make the decision in a depleted state.
• Block the next 48 hours from new commitments. “Don’t put anything optional on the calendar Wednesday or Thursday.” Recovery is the work; protect the work.
• Set a 48-hour rule on self-assessment. Your read on how the loop went is unreliable in the first 48 hours — too depleted, too emotionally raw. Write down what happened factually (round 1: this question, my approach was X) but don’t try to decide whether you “did well.” Wait for the recruiter or for 48 hours, whichever comes first.

The post-loop crash is real even when the loop went well. It is not diagnostic of how you performed. Treating it as recovery data — not performance data — is the right frame.

Common failure modes

Treating prep volume as the metric. “I did 200 LeetCode problems” is a comforting number. It does not predict loop performance. Energy distribution across the 14 days predicts loop performance. The metric to watch is can I sustain mid-day cognitive work the week of the loop?, not how many problems have I solved?

Skipping the taper because it feels lazy. The taper is the hardest tactic to execute because it triggers anxiety (“I should be studying”). The candidates who skip it routinely outperform during prep weeks and underperform during the loop. The taper is the strongest single predictor of round 4–5 performance.

Optimizing loop day without optimizing the day before. A perfect loop-day morning routine is undermined by a 6-hour T-1 study session. The best loop-day routine is the one that follows three nights of even sleep and a calm T-1.

Front-loading round 1. Burning 30% of your daily budget on round 1 because the adrenaline made you sharp. Round 1 is for showing the bar; the showcase happens in round 3 when most candidates have faded.

Ignoring the cumulative cost of a loop on the next two weeks. A real loop costs 3–7 days of recovery. Scheduling another loop the week after, or a major work project starting Monday, treats recovery as optional. It isn’t. Account for the recovery in your post-loop calendar before you accept the loop date.

Disclosing under stress instead of by plan. If you decide mid-loop that you’re depleted and need to ask for a break or an accommodation, your decision-making at that moment is impaired by the depletion you’re trying to address. Decide accommodations before the loop, not during it.

What to do next

This week:

• Log your real energy curve for three workdays (§1). One sketch, one page.
• Audit your next 14 days for cuttable drain (§3). Identify and cut at least three sources before adding any prep hours.

T-7:

• Convert the standard study plan into the energy-budget shape (§2). Move heavy lifting to T-14 to T-8; consolidate at T-7 to T-4; taper at T-3 to T-1.
• Decide on accommodations now (§7), not later. The decision in T-2 panic is the wrong decision.

T-3:

• Begin the taper (§4). One light review session per day, fixed sleep, reduced input volume.

T-0:

• Run the spend-evenly model (§5). Take the breaks. Eat once, protein-fat. Walk into round 1 calibrated, not maxed.

T+1:

• Block the next 48 hours. No self-assessment. Recovery is the work.

Related

• Guide: Surviving a Full Interview Loop with ADHD — prep systems and in-round mechanics. This piece runs on top of those systems; the energy budget is the resource the systems spend.
• Plan: Complete Senior Backend Interview Prep — the surface-by-surface 14-week plan. The taper structure here drops onto the final two weeks of that plan.
• Framework: The Behavioral Interview for Senior+ — behavioral rounds carry an outsized share of the loop’s working-memory cost. Pre-tagged stories reduce the cost.

Date-stamped: written 2026-05-31. The energy-budget model is stable. Specific accommodation policies vary by company and jurisdiction — verify current offerings with recruiters directly.