Google GCP production with Auto scaling for my simple ML python code
Budget: $30 – $250 USD
EASY PROJECT IF YOU KNOW GCP ML – SIMPLE ML CODE I GIVE YOU, YOU ONLY ASKED HOW TO DEPLOY
I give you base code and data or you can use your code and data if I like it
0.prove you did similar projects in past
1. ML project end to end development on Google GCP
2. Any python package installed by pip potentially can be used for this ML project – fully customised ML project capability
3. Build with Automatic scaling a customized model training for python code ML model provided by me or you can use your code if I like it
4. Build with Automatic scaling a customized inference - as more request as more VMs used and as less request as less VMs used (fast VMs number adjustment )
5. Automatic scaling, load balancing to handle 1234 requests/sec, requests done in parallel by python from any many computer on web
6. many VMs Distributed ML model Training with Hyperparameters search
7. many VMs Distributed Preprocessing (optional)
8. many VMs Distributed Inference
9. requests implemented by python running on windows computer. Predictions may be done from any computer on web no credentials needed (not ssh or any remote connection to GCP with credentials)
10. no GUI design, but all done by python SDK
11. basic ML code and data will be provided or you can use your code and data if I like it
YOU PROVIDE PYTHON CODE for GCP AND
VIDEO (my guess it will be around 60 minutes video) WITH EXPLANATIONS STEP BY STEP HOW TO DO
---------------------------------------------------------------
PROJECT ACCEPTANCE : I CAN DO BY MYSELF on my GCP account
----------------------------------------------------------------
TIME MAX 1 WEEK
=====================================================
Testing :
0. Do local computer prediction
1 For cloud prediction I use your python code which hosted on my GCP cloud project created by me using your instructions.
2 Make sure there are no dropped requests. Test in python code for requests if there are requests without correct response (calculate prediction for request locally and then compare with returned from GCP )
3 Many requests are sent in parallel from different computers: REQUEST consists of unique ID and textual data
4 Response: unique ID and prediction
5 When predictions are received: unique returned ID is used to match ID from GCP prediction and request text data local prediction.
ID to confirm it is the same request text used on GCP what was sent for prediction
6 compare local and remote prediction, calculate how many are the same
***************************************************************************
I do not give you access to my GCP account
make development on your GCP account
I give you base code and data or you can use your code and data if I like it
0.prove you did similar projects in past
1. ML project end to end development on Google GCP
2. Any python package installed by pip potentially can be used for this ML project – fully customised ML project capability
3. Build with Automatic scaling a customized model training for python code ML model provided by me or you can use your code if I like it
4. Build with Automatic scaling a customized inference - as more request as more VMs used and as less request as less VMs used (fast VMs number adjustment )
5. Automatic scaling, load balancing to handle 1234 requests/sec, requests done in parallel by python from any many computer on web
6. many VMs Distributed ML model Training with Hyperparameters search
7. many VMs Distributed Preprocessing (optional)
8. many VMs Distributed Inference
9. requests implemented by python running on windows computer. Predictions may be done from any computer on web no credentials needed (not ssh or any remote connection to GCP with credentials)
10. no GUI design, but all done by python SDK
11. basic ML code and data will be provided or you can use your code and data if I like it
YOU PROVIDE PYTHON CODE for GCP AND
VIDEO (my guess it will be around 60 minutes video) WITH EXPLANATIONS STEP BY STEP HOW TO DO
---------------------------------------------------------------
PROJECT ACCEPTANCE : I CAN DO BY MYSELF on my GCP account
----------------------------------------------------------------
TIME MAX 1 WEEK
=====================================================
Testing :
0. Do local computer prediction
1 For cloud prediction I use your python code which hosted on my GCP cloud project created by me using your instructions.
2 Make sure there are no dropped requests. Test in python code for requests if there are requests without correct response (calculate prediction for request locally and then compare with returned from GCP )
3 Many requests are sent in parallel from different computers: REQUEST consists of unique ID and textual data
4 Response: unique ID and prediction
5 When predictions are received: unique returned ID is used to match ID from GCP prediction and request text data local prediction.
ID to confirm it is the same request text used on GCP what was sent for prediction
6 compare local and remote prediction, calculate how many are the same
***************************************************************************
I do not give you access to my GCP account
make development on your GCP account