AI Engineer: How Delegating Code to Models is Changing the Developer's Role
38-year-old AI engineer Rohan Gor from marketing company Reach3 Insights has fully delegated code-writing tasks to artificial intelligence since December. Over months of experimentation, he sped up feature development 10–15 times, but ran into burnout and fears of losing his skills. Now the focus has shifted to architecture, research, and critiquing product decisions.
Shifting from Coding to Strategic Tasks
Gor graduated with a computer science degree in 2010 and has worked in marketing research ever since, where programming was a key part of his daily routine. Delegating code to AI has freed up time for higher-level tasks:
- Software architecture: designing systems without routine implementation.
- Research: in-depth analysis of solutions using AI queries.
- Product critique: evaluating project managers' decisions based on rapid prototyping.
AI can generate code in hours, not weeks. What used to take a month now gets done in 2–3 days. This allows for iterative hypothesis testing and adjustments.
The engineer stresses that performance expectations haven't dropped. AI handles the coding, but the workload keeps growing, leading to overload.
Benefits of AI in Day-to-Day Development
Delegating routine tasks changes the engineer's workflow. Instead of writing boilerplate code, Gor crafts prompts, reviews outputs, and integrates them.
Key advantages:
- Faster research: AI answers architecture questions, suggesting options with explanations.
- Less deadline pressure: more time for thinking and reviewing.
- Expanded role: from executor to co-creator of product decisions.
Example: When designing a feature, AI generates multiple prototypes. Gor evaluates them based on performance metrics and usability, picking the best one. This shortens the feedback loop from weeks to days.
Risks and Psychological Challenges
Despite the efficiency gains, Gor feels anxious. Programming is a skill honed over 15 years. Automating it raises questions about the future value of his expertise.
Main concerns:
- Burnout: more tasks while maintaining quality.
- Uncertainty: what if AI takes over architecture?
- Skill loss: coding skills atrophy without practice.
He notes that change is inevitable, but the transition requires adaptation. He recommends colleagues start with simple tasks and scale up gradually.
Key Takeaways
- Delegating code to AI speeds up development 10+ times, freeing time for architecture and research.
- Performance expectations remain the same, fueling burnout.
- AI expands the engineer's role to the product level, but sparks fears of losing expertise.
- The shift works best for middle/senior developers experienced in prompting.
- Future: focus on critique and design, with coding fading into the past.
— Editorial Team
No comments yet.