Rabib Alam is a Meta engineer and a former TPM at Tesla. The episode traces his path from Bangladesh to Meta AI.
In this episode Ahmed Fahad asks Rabib to explain data centers: the global backlash against AI and data centers, Meta's build-out with Hyperion and Prometheus, how a query reaches a data center, the difference between inference and training, and co-designing chips with Nvidia and AMD. The second half is advice for Bangladeshi students who want to land a job in Silicon Valley, in three parts, followed by career lessons from Tesla and Meta.
What the conversation covers
- Rabib's path from Bangladesh to Meta AI
- The global backlash against AI and data centers
- Meta's data center build-out: Hyperion and Prometheus
- How your query reaches a data center, and inference versus training
- Co-designing chips with Nvidia and AMD
- Advice for Bangladeshi students, part 1: believe in yourself
- Part 2: choosing your route, education or employment
- Getting noticed: the catered message and the "postal code" line
- Part 3: learning fast and finding your people
- Tesla, Meta and career advice on imposter syndrome
The conversation is in Bangla.