🤖 Top AI Interview Questions: Part-5
Advanced AI, NLP, and deployment concepts for interview preparation.
- Tokenization: Splitting text into units (words, sentences). Example: "I love AI" → ["I", "love", "AI"]
- Stemming: Reducing words to root form. Example: "running", "runner" → "run"
- Used in NLP for text preprocessing and normalization.
BERT (Bidirectional Encoder Representations from Transformers) is a transformer-based model.
- Reads text bidirectionally (context from both sides).
- Pre-trained on large corpus, fine-tuned for tasks.
- Use cases: Sentiment analysis, Question answering, Named entity recognition, Text classification.
- Focuses on relevant parts of input sequence when predicting.
- Captures relationships between words regardless of distance.
- Key component in BERT, GPT, T5; improves context understanding.
Predicts next word or sequence based on previous context.
Examples: GPT, BERT, LLaMA.
Used in: chatbots, autocomplete, summarization, translation.
YOLO (You Only Look Once) is a real-time object detection system.
- Processes image in one pass (single CNN).
- Outputs bounding boxes & class probabilities.
- Fast & efficient — suitable for real-time apps like surveillance & autonomous vehicles.
Makes AI decisions understandable to humans.
- Builds trust
- Useful in regulated industries (healthcare, finance)
- Techniques: SHAP, LIME, attention maps
- Interpretability: Ease of understanding model (esp. simple models).
- Explainability: Explaining decisions of complex models (e.g., deep learning) via tools.
- Both are key for trust, compliance, debugging.
Steps to deploy:
- Train and validate model
- Save model (Pickle, Joblib)
- Wrap in API (Flask, FastAPI)
- Containerize (Docker)
- Host on server/cloud (AWS, Heroku, Azure)
- Monitor performance & update regularly
- Bias & fairness
- Privacy & data security
- Job displacement
- Misinformation & deepfakes
- Lack of transparency
Crafting inputs to guide large language models (GPT, etc.) to produce desired outputs.
Techniques: few-shot, zero-shot, chain-of-thought prompting.
Critical for building AI apps, chatbots, and LLM tools.
Comments (0)
No comments yet
Be the first to share your thoughts!
Leave a Comment