DataTalks.Club Inside the AI Engineer Role: Tools, Skills, and Career Path - Ruslan Shchuchkin
Mar 20, 2026
Ruslan Shchuchkin, a GenAI engineer who builds production systems and side projects like Branch GPT. He discusses switching from business to AI engineering. Short takes on building in public, full-stack LLM products, using tools to move fast, starting local AI meetups, and why project track record and product discovery matter more than degrees.
AI Snips
Chapters
Transcript
Episode notes
Discover Product Value Before Optimizing Models
- Product discovery precedes core LLM optimization; validate user value before fine-tuning models.
- Ruslan iterates prompts, latency, cost, smaller models and fine-tuning only after confirming the use case with users.
Use Real User Sessions To Inform LLM Outputs
- Observe real users directly to discover features and structured outputs LLMs should produce.
- Ruslan used his own 'chat with your data' POC to learn what structured suggestions and next actions users actually need.
Filter News With Curated Sources And Meetups
- Use X (Twitter), Hacker News, and meetups as curated filters to keep current; prioritize trusted people over raw volume.
- Ruslan subscribes to X, uses Grok for aggregated SOTA, and asks peers at meetups which tools are battle-tested.