Hiver Computer Vision Engineer Interview: Rounds, Questions and Preparation

What the job involves

A Computer Vision Engineer in AI and deeptech builds systems that extract meaning from images and video, often for document understanding, OCR, object detection, or visual quality control. Day to day involves data collection and annotation, model training and evaluation, and deploying optimised models to production. At a company like Hiver, this could mean automating extraction from emails and attachments or classifying visual content at scale.

Role and Technical Knowledge

1. 1

   Explain how a convolutional layer works and why it is useful for image data.

   Cover local receptive fields, weight sharing, and translation invariance. Mention how pooling reduces spatial dimensions.

2. 2

   What is the difference between object detection and image segmentation?

   Detection gives bounding boxes and class labels; segmentation gives pixel-level masks. Give a use case for each, like counting cars vs outlining tumours.

3. 3

   How would you handle a dataset with very few labelled images for a custom vision task?

   Discuss transfer learning, data augmentation, semi-supervised learning, and synthetic data. Mention starting with a pre-trained backbone like ResNet.

4. 4

   What are some common evaluation metrics for object detection and when would you use each?

   Explain mAP, IoU, precision-recall curves. Note that mAP is standard for detection, while IoU is used for segmentation overlap.

Practical Scenarios

1. 5

   You need to extract text from scanned invoices with varying layouts. How would you approach it?

   Pipeline: preprocessing (deskew, denoise), OCR (Tesseract or a deep model), then layout analysis or key-value extraction. Mention handling rotation and low contrast.

2. 6

   Your model performs well on validation but poorly in production. What could be wrong and how do you debug it?

   Check for data drift, preprocessing mismatch, latency issues, or overfitting. Suggest monitoring, logging, and A/B testing.

3. 7

   How would you deploy a computer vision model to run in real time on a CPU-only server?

   Mention model compression (quantisation, pruning), using ONNX Runtime or OpenVINO, batching, and choosing a lightweight architecture like MobileNet.

4. 8

   You have to build a system to detect whether a document contains a signature. What steps would you take?

   Start with data collection and annotation, then a classification or detection model. Discuss class imbalance and false positive cost.

Behavioural and Project Questions

1. 9

   Tell me about a computer vision project you built. What was your role and what was the outcome?

   Use STAR format. Highlight the problem, your technical choices, challenges, and measurable results like accuracy or speed improvement.

2. 10

   Describe a time you had to learn a new tool or technique quickly for a project.

   Show initiative and resourcefulness. Mention specific resources, time taken, and how it helped the project.

3. 11

   How do you ensure your models are fair and unbiased, especially when working with human-centric images?

   Discuss diverse datasets, bias audits, and evaluating performance across subgroups. Mention ethical considerations in deployment.

4. 12

   Give an example of a technical disagreement with a teammate and how you resolved it.

   Focus on listening, data-driven discussion, and compromise. Show you value team success over being right.

HR and Company Fit

1. 13

   Why do you want to work as a Computer Vision Engineer in a product company like Hiver?

   Connect your interest in applied AI to solving real business problems. Mention Hiver's focus on customer service and email collaboration.

2. 14

   Where do you see yourself in three years in this role?

   Show ambition to grow into a senior engineer or tech lead, with deeper expertise in vision and deployment. Align with company growth.

3. 15

   How do you keep up with the fast pace of AI research?

   Mention following arXiv, attending webinars, reading blogs, and implementing papers. Give a recent example you found interesting.

4. 16

   What are your salary expectations for this role?

   Research typical fresher salaries for computer vision roles in India. Give a range and express flexibility based on overall compensation.