Interesting perspective Amit. When I see your saas engineer vs ai engineer comparison, I read it more as feature‑factory vs product‑/experiment‑driven teams than SaaS vs AI. Is there anything here you’d consider truly specific to AI engineering, versus good product engineering in general?
You are right, Ajitha, they are very similar. The underlying models are so good now that building AI products is rarely about core AI techniques. Those stay below the level of abstraction where we operate. The real shift is who does the work. In the SaaS era, a human drove a tool to get an outcome. Now, the expectation is that the agent achieves the outcome directly. That means the agent has to handle the messiness of the real world, unpredictable corner cases, and subjective judgment. That is exactly why a data-driven, experimental approach is mandatory now, rather than just shipping well-defined functionality as we did in SaaS.
Interesting perspective Amit. When I see your saas engineer vs ai engineer comparison, I read it more as feature‑factory vs product‑/experiment‑driven teams than SaaS vs AI. Is there anything here you’d consider truly specific to AI engineering, versus good product engineering in general?
You are right, Ajitha, they are very similar. The underlying models are so good now that building AI products is rarely about core AI techniques. Those stay below the level of abstraction where we operate. The real shift is who does the work. In the SaaS era, a human drove a tool to get an outcome. Now, the expectation is that the agent achieves the outcome directly. That means the agent has to handle the messiness of the real world, unpredictable corner cases, and subjective judgment. That is exactly why a data-driven, experimental approach is mandatory now, rather than just shipping well-defined functionality as we did in SaaS.