Target Companies — Prep Guide
Company-specific prep begins Week 21–22, but read this early to guide your project choices.
🟣 Anthropic
Interview Style: Deep AI + safety thinking + practical coding
Must Know:
- Constitutional AI (read the paper summary)
- RLHF and how Claude is trained
- Responsible Scaling Policy (RSP)
- How Claude differs from GPT architecturally
- Claude's Character and values
Key Questions They Ask:
- "How would you make an AI system safer?"
- "What do you think about AI x-risk?"
- "Walk me through Constitutional AI"
- "How would you evaluate if a model is safe?"
Project to Build: Something demonstrating safety-aware AI design
🟢 OpenAI
Interview Style: Agents, tools, fast iteration, capability + safety balance
Must Know:
- GPT architecture basics (transformer, attention)
- RLHF and fine-tuning process
- Multimodal models (GPT-4V, Whisper, DALL-E)
- Reasoning models (o1, o3 — how they differ)
- OpenAI API: assistants, tools, structured output
Projects they love: Agents, tool use, creative applications
🔵 Google / DeepMind
Interview Style: Heavy LeetCode + system design at Google scale
Must Know:
- Solve Hard LeetCode consistently
- Gemini architecture and products
- Vertex AI, Google Cloud AI stack
- "Design at Google scale" (billions of users)
- MapReduce, Bigtable mental models
Key Difference: Google has the hardest DSA bar of all target companies.
🔵 Meta AI
Interview Style: DSA + open-source AI knowledge
Must Know:
- Llama model family (Llama 2, Llama 3)
- PyTorch fundamentals
- FAISS (Facebook's vector search library)
- Open-source AI philosophy
- Meta's AI research areas (FAIR)
Values: Open source, fast iteration, scale
🟡 Sarvam AI
Interview Style: Practical AI + passion for India AI ecosystem
Must Know:
- Indic language AI challenges
- Low-resource NLP techniques
- Multilingual model architectures
- India's AI landscape and opportunities
Project to Build: Anything involving multilingual/Indic language processing
🟠 PayPal
Interview Style: Applied AI + software dev combo
Must Know:
- AI in fintech — fraud detection patterns
- Risk models and compliance AI
- Safety and guardrails for financial AI
- Combine your AI + software dev strength
Key Angle: Your software dev background + AI knowledge is a perfect fit
⚪ Syncfusion
Interview Style: Practical AI for developer tools
Must Know:
- AI-enhanced UI components
- Code generation for components
- Developer tools and workflows
Project to Build: Something with Syncfusion products + Claude API
📋 Company Research Tracker
| Company | AI Products Known | Key Papers Read | Mock Done | Ready? |
|---|---|---|---|---|
| Anthropic | ⭕ | ⭕ | ⭕ | ❌ |
| OpenAI | ⭕ | ⭕ | ⭕ | ❌ |
| ⭕ | ⭕ | ⭕ | ❌ | |
| Meta | ⭕ | ⭕ | ⭕ | ❌ |
| Sarvam AI | ⭕ | ⭕ | ⭕ | ❌ |
| PayPal | ⭕ | ⭕ | ⭕ | ❌ |
| Syncfusion | ⭕ | ⭕ | ⭕ | ❌ |