8 ms·

Thanks, I've been trying to be as proactive as I can but I could probably do more. A big issue is that the feature in question sounds simple but is really quite

Thanks, I've been trying to be as proactive as I can but I could probably do more. A big issue is that the feature in question sounds simple but is really quite complex (e.g. tell whether a picture is of a bird) and I'm not sure I'm conveying this effectively to the CEO. It's really a confluence of problems: lack of communication due to high friction, lack of organization, lack of documentation, and lack of mock-ups or user stories. What I'll try to do is create and assign myself tasks on our task tracker so we all at least have some idea if what I should be doing, and I can use this as a starting point to ask for clarification.

throwawayReply 10y ago
Don't expect others to produce user stories or mock-ups, especially if this isn't already part of the culture. If you want to have user stories and mock-ups, start spending your time making those yourself and push them back to whoever is giving you the requirements for comment. If they question the need for you to be spending time doing that, it's an opportunity to ask for more specific direction on what you should be spending your time doing. Chances are it's a classic case of not knowing what they want, but knowing what they don't want, so by having their requirements reflected back to them they'll come back with "Ah, but what I meant was..." and you'll get more clear requirements. If it's still murky after that and you really can't get more concrete requirements than "We need this magic box made" then consider that some places genuinely just aren't good places to work, and consider this a learning opportunity and point for future comparison.
abannin 10y ago
I suspect the issue is one of two scenarios: 1) CEO has low bandwidth, hands off a nice-to-have-but-not-critical feature. They don't have to invest much time; if it works, awesome, if not, no problem. If you don't complete the feature, it may or may not get done. 2) Company is overwhelmed, feature is perceived to be easy(-ish). It will get done, you're just a first pass. If (1), you're probably best served not asking for much help. If (2), you need to educate & manage expectations about the complexity and necessary steps for the feature. Either case, you would probably be best served by prioritizing the input/output functionality (input a picture, output a prediction) over optimizing precision/recall on the algo. This allows you to 1) show progress to others and 2) you can start having conversations about false positives/negatives by showing different images and the prediction output.

throwawayReply 10y ago

Don't expect others to produce user stories or mock-ups, especially if this isn't already part of the culture. If you want to have user stories and mock-ups, start spending your time making those yourself and push them back to whoever is giving you the requirements for comment. If they question the need for you to be spending time doing that, it's an opportunity to ask for more specific direction on what you should be spending your time doing. Chances are it's a classic case of not knowing what they want, but knowing what they don't want, so by having their requirements reflected back to them they'll come back with "Ah, but what I meant was..." and you'll get more clear requirements. If it's still murky after that and you really can't get more concrete requirements than "We need this magic box made" then consider that some places genuinely just aren't good places to work, and consider this a learning opportunity and point for future comparison.

abannin 10y ago

I suspect the issue is one of two scenarios: 1) CEO has low bandwidth, hands off a nice-to-have-but-not-critical feature. They don't have to invest much time; if it works, awesome, if not, no problem. If you don't complete the feature, it may or may not get done. 2) Company is overwhelmed, feature is perceived to be easy(-ish). It will get done, you're just a first pass. If (1), you're probably best served not asking for much help. If (2), you need to educate & manage expectations about the complexity and necessary steps for the feature. Either case, you would probably be best served by prioritizing the input/output functionality (input a picture, output a prediction) over optimizing precision/recall on the algo. This allows you to 1) show progress to others and 2) you can start having conversations about false positives/negatives by showing different images and the prediction output.