Search Jobs
Contract: Data Integration Architect / South East London
This job has expired or may no longer be taking applications, but other similar jobs are available.
Added: | 2022-04-06 |
Location: | South East London |
Salary: | Good |
Duration: | Contract |
Apjid | 3 |
Data Integration Architect – 6 Months Contract – London/Remote
Required:- Build Advanced Regression Machine Learning Algorithm
The client is working in the technology field and have experimental data collected from 335 sensors. Each experiment contains X-values with their time steps and Y-values. X is the input (predictor) variable. Y is the target (response) variable that represents the true values for any experiment.
The main objective of this project is to build an effective machine learning (ML) algorithm to predict Y from X for any experiment. The developed ML must account for the baseline deviations for the sensors and inherited noise. My client wants to use deep learning algorithms and advanced regression algorithms.
Data Integration Architect Requirements:
Notes:
Candidates must be eligible to work in the UK
Data Integration Architect – 6 Months Contract – London/Remote
Required:- Build Advanced Regression Machine Learning Algorithm
The client is working in the technology field and have experimental data collected from 335 sensors. Each experiment contains X-values with their time steps and Y-values. X is the input (predictor) variable. Y is the target (response) variable that represents the true values for any experiment.
The main objective of this project is to build an effective machine learning (ML) algorithm to predict Y from X for any experiment. The developed ML must account for the baseline deviations for the sensors and inherited noise. My client wants to use deep learning algorithms and advanced regression algorithms.
Data Integration Architect Requirements:
- The developed model will be applied in real-time applications, meaning X-values are applied to the model at each time step so consider that the model must predict at each time step.
- The final ML algorithm will be deployed as an android APK so consider the complexity of the proposed model and its computational time.
- The root mean square error (RMSE) metric must be used to assess the performance of the model. The target is to get RMSE below 1.4 using the testing data not using the training data.
- My client has randomly split the data into 80% for training and 20% for testing. The performance of the model must be evaluated using the testing data to avoid overfitting and data leakage.
- It is not acceptable to use an id for the sensor datasets this will be considered data leakage.
- The ML algorithm must be implemented using MATLAB or Python.
Notes:
- X-values contaminated with sensor noise, drift, and baseline deviations. Noise removal and detrend the sensor data is required before making the prediction
- The client can use the first 24 observed values to calculate useful features such as the mean, median and standard deviation etc. In real-time application, we can pause the model for up to 24 readings to calculate these features and then update, tune and adjust the model parameters if required.
- The client cannot clean and detrend the data manually prior to the prediction. This should be done within the model implementation.
Candidates must be eligible to work in the UK
Data Integration Architect – 6 Months Contract – London/Remote
How to Apply
© All rights reserved, 2001 - 2024