Let's say you only have resources to complete 6 of 10 tasks you're allocated at Meta. How would you prioritize?

I would use  a framework like RICE, Moscow, Story mapping etc. to prioritize the tasks with my stakeholders. Get high confidence estimates and then decide how many of the 10 tasks can be completed on time with the available resources. The tasks with prioritized based on Reach * Impact * Confidence/Effort. Highest score is the highest priority. I would also make sure there is RAID (Risk, Assumption, Issues and Dependency) log for the tasks that make the cut. Risks should have mitigation plan and dependencies should be committed to be delivered in time to minimize risk to schedule. The tasks that are on the critical path and should be scheduled when the dependencies are available.

If I had only enough resources to complete 6 tasks out of 10 I would first prioritize the tasks based on the urgency of each task and rank them in order of which task needs to be delivered first. Then I would allocate my resources to finish the jobs which are common to all tasks if any. I would allocate my resources in a way that I can maximize the profitability of the whole project for example focus on delivering a certain feature that is famous and helpful to the masses and hence improve scalability. I would also check for dependencies of each task on the other which means I would allot maximum resources to a task which when completed could ease the completion of the other tasks automatically.

For example while I was working at Alpha Design Technologies we had only 2 Spectrum analysers, 2Oscilloscopes that could be used for testing and 2 skilled assembly and rework technicians in-house who had to complete the assembly of 5 different products, so once the PCBs arrived from Manufacturing I made sure that the Power supply unit was the first PCB to be assembled and up and running because all other modules were dependent on the power supply PCB for their power. the next one would be the Transmitter PCB as it is a very complex assembly with high number of components and requires maximum testing time. I also discussed with my manager that it would be best if we outsource the service of Assembling multiple PCBs so that we would be able to get multiple PCBS assembled simultaneously and they would be ready to be tested once delivered back to us.

Understand the following for each task:

* Impact of the task on a given product. Is it absolutely essential for a minimum viable product.

* When the task needs to be delivered based on the product timeline

* Quantify effort vs benefit -- low effort, high benefit tasks are more valuable.

* Discuss with the stakeholders on the importance of the task -- are there alternatives?

* Based on the above finding put a plan together to deliver the 6 tasks, show the when remaining 4 could you be delivered. Present the plans to the stakeholders to get buy in.

Would start with the tasks that have the lowest resource needed with highest ROI and work our way down. I would also look to balance between the tasks that are not wasteful work for later, have better long term ROI and help parallelize what can be accomplished in the given time