NVIDIA面试 复盘来了！包括以下来自成功上岸的 x 同学的分享：
写在最前面，感觉 NVIDIA 还是非常注重 match 程度，面试里的每一个人都问了我同样的几个问题：我喜欢做什么、我想做什么。
一面：Manager Phone Interview（45min）
- 自我介绍，问专业课程及项目经历。
- Coding 概念问答 + 简单题目。
- 组里工作介绍 & 行为面试。
二面：Engineer On-site 电面（45-60min）
- Container、DevOps、软件开发流程、命名原则等细节问答。
- Coding + C++/算法 + System Design 题量较大。
Nvidia Interview Questions
- In your employment experience, what suggestions have you proposed for your supervisor?
- Do you have any coding experience with assembly code?
- What is your most challenging project experience?
- What parallel programming experiences do you have? MPI? CUDA? Multi-threading?
- How are you good at Python and C++? Describe related study and project experiences.
- Why would you like to participate in this internship?
- Any CUDA programming experiences?
- What do you know about NVIDIA’s GPU architectures (Fermi, Turing, Ampere)?
- Tell me your most difficult and fulfilling project experience.
- Which algorithm-related courses did you take?
Nvidia Coding Question 1
构造序列：从 1 开始，每个整数 i 重复出现 i 次，直到无限。 给定若干 queries（索引列表），返回对应元素。
# Python 示例
def generate_list(n):
lst, cur = [], 1
while len(lst) < n:
lst.extend([cur] * cur)
cur += 1
return lst[:n]
def query_elements(lst, queries):
return [lst[q] for q in queries]
queries = [2, 3, 6]
lst = generate_list(100)
print(query_elements(lst, queries)) # [3, 3, 5]
Nvidia Coding Question 2
对数组 arr，可交换相邻同奇偶性的元素。 目标：拼接成字符串（例如 [1,2,4,3] → “1243”），求最大可能值。
# Python 示例
def max_number_from_array(arr):
odd = sorted([x for x in arr if x%2], reverse=True)
even = sorted([x for x in arr if x%2==0], reverse=True)
o, e, res = 0, 0, []
for x in arr:
if x%2==0:
res.append(even[e]); e += 1
else:
res.append(odd[o]); o += 1
return ''.join(map(str, res))
print(max_number_from_array([1,2,4,3])) # "3421"
Read More
Contact Us
经过我们的面试辅助和 OA 代写，候选人不仅展示了编程能力，也体现了清晰思路与有效沟通。 这些经验不仅有助于 Nvidia 面试，也能提升解决实际问题的能力。祝大家面试顺利！
如果你也需要我们的面试辅助服务，请 立即联系我们。
END
