I am employed as developer in what was a Startup once, now going into its maturity.

There is a huge project coming up, probably blowing the budget 10-fold. Budget was 1-digit million before. Now of course, management wants this to happen. I as a developer have doubts.

My employer has licensed some kind of properitary algorithm, which does stuff. I was employed to develop a framework to allow said stuff to happen in parallel, and make scale happen. For years, I was sole developer, so a lot of the foundations bear my name.

Now this was a Startup, short on manpower, so i took other jobs. Like the job of fixing said algorithm to actually work. It was developed for a similar but not so demanding problem, in its own right working properly, but for the current problem some basic assumptions do not hold. We don't have any sources available, just binaries. So I cannot work on the algorithm itself.

I reckon it was a bad idea to take responsibility in the first place, but well, I did.

Now I am in a bad position, where there's a Major project (it got news coverage in a major western EU state) that I think is going 90% odds of failure, and I can either bail or try to make it happen. And...I am no Genius.

In the last year, we did a "prepation study" to make sure our Algorithm is capable of succesfully solving the problem. We were given "Dummy" material which was supposedly simulating the case, making me think it was all going well. It did not. Early this year, we got our first sight on the real stuff, and it was - as said - breaking basic assumptions I cannot fix because we dont have any sources. Surpise, work it did not. So in our presentation to our client, we used "fake" stuff. "Fake" as in we used knowledge, that in the real case, will not be available. Think about it as if you knew the ground truth to a classification problem and use that to solve it. It worked because the sample set was limited and we could prepare it. That will not work in the real use case.

This thing has news coverage spanning my country. I am worried about my reputation. There is nothing I can do to make it work; There are people who could, but I dont know if it fits their timetable and bills (likely, it wont).

I do not wish to hurt my employer as in leaving in a hurry. Right now, the project is im prepation. I am not involved in any ongoing projects. I am thinking of quitting with a Years notice, combined with a statement that I will no longer assume any responsibilities with said project (There are more I could spend my time with, plus writing docs.).

Stay or Run?

  • 1
    Is there anyone mature/senior /responsible with whom you can have a frank discussion ? Instead of running , if you present the things you said , the person would definitely weigh your thoughts and take steps to resolve the issues ( may involve you too ). Worst case you can run any time. Commented Mar 30, 2016 at 2:43
  • "I am thinking of quitting with a Years notice, combined with a statement that I will no longer assume any responsibilities with said project" - If you do decide to get out of there, don't give a year's notice (too long) and don't say things about "no longer assume any responsibilities with said project". Simply give the notice your contract requires, or enough notice to allow a handover of the remaining responsibilities to another person. That will ensuer your employer is not left in a bad position.
    – Brandin
    Commented Mar 30, 2016 at 8:25
  • 4
    "Stay or run?" is asking us to make up your mind for you without knowing all of the details. That's not only irresponsible on your part, it's impossible for us. This site is about questions with practical answers that can benefit more than just one person which makes your question off-topic and unsuitable.
    – Lilienthal
    Commented Mar 30, 2016 at 8:56
  • Why a year's notice? It's not your problem, just get a new job and quit. Commented Mar 30, 2016 at 13:24

5 Answers 5


"So in our presentation to our client, we used "fake" stuff"... Extremely bad idea, for the reasons you have discovered. Better to tell the customer there's an issue and give an estimate of how long it will take to fix.

If you can't deliver on time, the customer would rather hear that while there is still time to adjust their own plans. Delaying costs them much more money.

Either convince yourselves that this is fixable or warn the customer. Preferably both.

  • sometimes you don't need a perfect working product for presentations, just a sort of proof of concept of part of an idea
    – Kilisi
    Commented Mar 30, 2016 at 6:15
  • "So in our presentation to our client, we used "fake" stuff"... Extremely bad idea" - well BILL GATES did that to IBM, so yeah, just FAKE it, but make sure you'll have that product later, or else you're pretty much f*cked, I wouldn't do that, unless I needed to xD, JK, I wouldn't ;)
    – Kyle
    Commented Mar 30, 2016 at 11:45
  • 1
    When giving a demo to potential clients, its always best to work with pre-prepared data and a carefully orchestrated script. Deviate from the script and the demo gods (or perhaps demo demons) will strike. Having an abundance of C-level people or venture capitalists in the room and the demo gods and demo demons will look for any and every possible avenue to strike. Commented Mar 30, 2016 at 11:46
  • I have faked demos, good purposes of producing a videotape in a rush -- but only when I knew we had the demonstrated capability working but temporarily out of service. Never promise what you don't think you can deliver; never demo what you don't __know __ you can deliver... (And don't get me started on Gates. He's an amazingly good salesman, I'll try to leave it at that.)
    – keshlam
    Commented Mar 30, 2016 at 14:41

You didn't say what your software does or how this proprietary software helps you accomplish that. Perhaps it doesn't work, perhaps it never worked, perhaps you deluded yourselves that it does work, perhaps you committed a massive hoax. Or perhaps there is something real behind the curtain and it just needs a bit of work. You and your employer should know which is the case.

Assuming there is something real behind the curtain, you now have 10 times your previous budget to make it work. Use that! Perhaps your employer could buy a part of the company that owns that software (so now its theirs). Or it could buy that proprietary software outright. Or it could purchase the rights to see and modify the source code to that proprietary software. Or it could buy a better license that provides support. Or it could hire someone (or more than one someone) who can recreate the algorithm. Reverse engineering is a lot easier than is starting from scratch.


You have a budget 10 times more than a single figure one in the millions? Make it work, if you can't, subcontract to someone who can.

Work out what you need, then find someone who can make it happen.


I'm a technical supporter of a 3rd-party API. We constantly deal with this situation, but from the other side. Our customers buy expensive license and find out that it does not work with the real-world samples. Here is what we do in cases like your one: we recommend our Professional Service. If you contact the 3rd-party and get their estimation for making the library work with your samples, you may find out, that it's not that bad.


It's your bosses responsibility if they enter a contract, and are paid cash in advance, to make it happen. Not yours. If they have a budget over ten million and they try to make this happen with you as the sole developer without access to source code, that is just stupid, and not your fault.

On the other hand, as long as they pay you, they pay you. Tell them the facts, that this isn't going to work without either major support by your supplier, or by getting a license to their source code that allows you using it and changing it in whatever way you like (if the company has a ten million budget, that shouldn't be a problem), and by adding more experienced man power. That's all you can do.

So as long as they pay you, you are fine. Since doing overtime won't fix the company's problem, don't let yourself be pressured into doing overtime. When you get a feeling that your salary payments are getting unsafe, you look for a new job. Or you look earlier, if you've had enough of it.

Giving a year's notice is nonsense. Ask yourself: How does this benefit you, compared to working there until you think it's time to leave now, and giving two weeks notice or the minimum that is required? And "how does it benefit you is the only question you need to ask.

Rejecting any responsibility is also nonsense. It's not your responsibility anyway. If your company asks you to do something that is impossible, that's their responsibility, not yours. Your responsibility is to do a job that is compatible in quality with your salary. If they think you're not doing that they need to find someone else. If doing a job that is compatible in quality with your salary isn't enough, then they need to find someone who is better and more expensive. But it's not your responsibility.

Not the answer you're looking for? Browse other questions tagged .