← Back to blog Amazon NG VO Loop 2026 Complete Debrief｜Post-OA 4-Round Onsite + 16 LP Behavioral Cheatsheet
VO Amazon

Amazon NG VO Loop 2026 Complete Debrief｜Post-OA 4-Round Onsite + 16 LP Behavioral Cheatsheet

2026-05-16

Amazon's NG OA is famously "loose" — 25-35% pass rate, and 1point3acres feedback consistently says "both Coding ACs + clean Work Style Survey" is enough. The real filter is the post-OA VO Loop: 4 rounds with Bar Raiser veto power, and the 2026 cycle overall offer rate is ~20%.

This article focuses on the post-OA onsite loop and skips the OA content (see our Amazon NG OA 2026 guide for that). Below, we break down Coding 1+2 + System/LLD + Bar Raiser BQ — real questions, scoring, and the 16 LP mapping.

Amazon NG VO Loop Overview

Round Duration Content Interviewer
1. Coding Round 1 60 min 1 LC Medium + 2-3 LP questions Senior SDE I
2. Coding Round 2 60 min 1 LC Medium-Hard + 2-3 LP Senior SDE II
3. System / LLD Round 60 min OOD design (NG usually LLD) + 2 LP Tech Lead / Senior
4. Bar Raiser 60 min 0-1 light coding + 5-6 deep LPs Bar Raiser (no hiring-team incentive)

2026 cycle changes:

Coding Round 1 + 2: Typical Real Questions

Q1 — Server Cluster Health Monitor (Coding 1, most frequent in 2026 spring)

Prompt: heartbeat time series for n servers. Each server emits one heartbeat per second. Implement:

Trap: n can be 1e6, naive scan TLEs.

from collections import defaultdict
from sortedcontainers import SortedList

class HealthMonitor:
    def __init__(self):
        self.last_seen = {}             # server_id -> latest ts
        self.by_ts = SortedList()       # (last_ts, server_id) for fast filtering

    def register(self, sid, ts):
        if sid in self.last_seen:
            self.by_ts.remove((self.last_seen[sid], sid))
        self.last_seen[sid] = ts
        self.by_ts.add((ts, sid))

    def failed_servers(self, now, k):
        threshold = now - k
        failed = [sid for sid, ts in self.last_seen.items() if ts < threshold]
        return sorted(failed)

Interviewer follow-ups:

Q2 — Order Batching with Constraints (Coding 2, harder)

Prompt: n orders, each with (deadline, weight). A batch can hold at most W total weight, and all orders in the batch must have deadline ≥ batch completion time (each batch takes fixed time t). Find the minimum number of batches.

Approach: greedy + priority queue — process by ascending deadline; pack while feasible, otherwise open a new batch.

def min_batches(orders, W, t):
    orders.sort(key=lambda x: x[0])  # by deadline
    count = 0
    current_time = 0
    cur_weight = 0
    for d, w in orders:
        if cur_weight + w > W or d < current_time + t:
            if cur_weight > 0:
                count += 1
                current_time += t
            cur_weight = w
        else:
            cur_weight += w
    if cur_weight > 0:
        count += 1
    return count

LP tie-in: each Coding round ends with 5-10 minutes on 1-2 LP stories (most common: Customer Obsession and Deliver Results). Don't burn all 50 minutes on coding — reserve 10 min for LP.

System / LLD Round: Typical OOD

High-Frequency Real Q: Simplified Amazon Parcel Locker System

Required APIs:

Scoring Dimensions

Dimension Penalty Bonus
Class design Everything in LockerSystem Split into Parcel / Locker / LockerBank / LockerService
Data structure List scan Size-class buckets (S/M/L/XL), each with a free queue
Concurrency Not mentioned Fine-grained lock per LockerBank for assign_locker and pick_up
Failure handling Not mentioned 3 wrong codes → notify support; locker full → suggest alternate site
Scalability "I'd use a database" Concrete schema: locker, parcel, event_log (the latter is the fact table)

Time Allocation (60 min)

0-5 min   Clarify requirements + clarifying questions
5-15 min  Core data model (class diagram + DB schema)
15-30 min API design + key algorithms (assign / pickup / metrics)
30-45 min Concurrency / scaling / failures
45-50 min Observability / testing strategy
50-60 min 2 LP questions (always)

Bar Raiser BQ: 16 LP Mapping

Amazon expanded LPs from 14 to 16 in 2024 (adding Strive to be Earth's Best Employer and Success and Scale Bring Broad Responsibility). Bar Raiser typically deep-dives 5-6 of them.

16 LPs × Story Direction

LP Story Direction High-Frequency Opener
Customer Obsession Pivot driven by customer feedback "Tell me a time you used customer feedback to improve a product."
Ownership Took on a mess project "Describe a time you took on something that wasn't your job."
Invent and Simplify Novel solution to an old problem "Walk me through a time you proposed a simpler solution."
Are Right, A Lot Disagreed with team and were right "Tell me about a decision others disagreed with."
Learn and Be Curious Picked up a new tech stack "How do you keep learning?"
Hire and Develop the Best Mentored juniors "How have you helped others grow?"
Insist on Highest Standards Pushed quality "Tell me about a time you raised the bar."
Think Big Expanded scope "When did you advocate for something ambitious?"
Bias for Action Decided fast with incomplete data "Tell me about a time you took action with incomplete data."
Frugality Did more with less "How have you done more with less?"
Earn Trust Repaired broken trust "Tell me about regaining someone's trust."
Dive Deep Root-cause investigation "Tell me about a complex problem you investigated."
Have Backbone; Disagree and Commit Pushed back, executed anyway "Describe a time you disagreed with your manager."
Deliver Results Delivered under pressure "Tell me about a high-pressure deadline."
Strive to be Earth's Best Employer ⭐NEW Improved team experience / DEI "How have you helped foster team well-being?"
Success and Scale Bring Broad Responsibility ⭐NEW Took on broader responsibility "How do you think about your work's broader impact?"

Bar Raiser Scoring Formula (reverse-engineered from 1point3acres data)

Final score ≈
   0.35 × LP fit (do you have stories for 8+ LPs)
 + 0.25 × Story specificity (concrete metrics / decisions)
 + 0.20 × STAR completeness
 + 0.10 × Quality of reverse questions
 + 0.10 × Communication (cadence, confidence, not arrogant)

Strong recommendation: write 12-16 STAR stories before onsite, with each story mapping to at least 2 LPs, so any LP can be answered with 1-2 stories.

Comp Range (2026 NG L4 SDE)

City Base RSU (4yr 5/15/40/40) Sign-on Y1 Sign-on Y2 Y1 Total
Seattle / Bellevue $170-180K $130-160K $30-50K $25-40K ~$210-250K
NYC / Bay Area $180-195K $140-170K $35-55K $30-45K ~$225-265K
Austin / DC $160-175K $120-150K $25-45K $20-35K ~$195-235K

Notes:

Timeline (OA Pass → Offer)

Week 0   OA pass email
Week 1-2 Recruiter call + onsite scheduling
Week 3-4 Onsite Loop (4 rounds in one day, all Zoom)
Week 5-6 Hiring Manager calibration + writing the hiring doc
Week 7-8 Offer letter / negotiation
Week 9   Accept (deadline typically 7-14 days)

Timing alert: Amazon's offer deadline is 7-14 days, shorter than Google / Meta (21+). If you have other ongoing loops, push Amazon's timeline back 2 weeks before onsite.

FAQ

Q1: What's the actual Amazon NG VO pass rate?

Post-OA VO pass rate is ~20-25%, more candidate-friendly than Google / Meta NG (10-15%). But Bar Raiser veto is real — 4-round strong-hires can still be rejected if Bar Raiser says no-hire. Bar Raiser carries ~35-40% of the loop weight.

Q2: Must I prep all 16 LPs for Amazon NG VO?

At least 12. Bar Raiser deep-dives 5-6; hiring-team interviewers each pick 2-3, so you'll be hit on 12+ across the loop. You don't need a unique story per LP — mapping one story to 2-3 LPs is standard practice (e.g., "rewrote legacy code" covers Ownership + Insist on Highest Standards + Dive Deep).

Q3: What language for Amazon NG VO coding?

Python / Java / C++ are the safe trio, with Python the NG default. JavaScript / Go: not forbidden but risky — Amazon's internal code is mostly Java, and interviewers may follow up in ways that expose unfamiliarity. Use what you know best — don't switch for culture fit.

Q4: How does Bar Raiser actually score?

A Bar Raiser is a senior engineer with no incentive in your hiring team — they don't gain from filling headcount. Scoring strictly aligns with the 16 LPs, and they only hire candidates who "raise the team's bar". A Bar Raiser no-hire almost always blocks the offer, even when the hiring manager wants you.

Q5: How long is the Amazon NG VO failure cooldown?

6 months — shorter than Google / Meta (12 months). But Amazon keeps an internal record, and Bar Raiser will be deliberately stricter on attempt #2. Strongly recommend submitting a "reflection letter" before re-applying — recruiter appends it to your packet.

Q6: NG L4 vs L5 — how is the level decided?

NG defaults to L4 (SDE I). Upgrading to L5 (SDE II) requires demonstrating "L5 behavior" during the loop: independent ownership in System Design / Bar Raiser, multi-team collaboration history, scope larger than a single feature. Most NGs land L4, but a Bar Raiser strong-hire occasionally triggers an "upgrade conversation".

Contact

If you're prepping Amazon NG / SDE II Loop, coding is not the variable — Bar Raiser BQ is where 70% of candidates fall short. We've curated 2025-2026 cycle Bar Raiser top-60 questions + 16 LP × story-mapping templates + STAR writing workshop.

Add WeChat Coding0201, get the Amazon Loop bank and Bar Raiser mocks.