Name: Esri Data Science Challenge 2019 at HackerEarth
Start: 2019-03-01T06:30:00
End: 2019-03-10T13:25:00
Location: HackerEarth

ANNOUNCEMENTS

The shortlists have been announced. Here's the shortlist criteria -

2019 batch/1-10 YOE
Scored above cut-off

Please note that this shortlist doesn't guarantee ...

more

ABOUT CHALLENGE

Esri is the world leader in geographic information systems (GIS) and developer of ArcGIS, the leading mapping and analytics software used in 75 percent of Fortune 500 companies. They are looking forward to hiring candidates for the profile of Data Scientist for the Esri R&D Center - New Delhi location.

Are you passionate about applying AI and machine learning to solve some of the world’s biggest challenges? So is Esri! At the Esri R&D Center in New Delhi, they are applying cutting-edge deep learning techniques to revolutionize geospatial analysis and derive insight from imagery and location data.

Participate in this challenge to get a shot at joining their team of exceptional data scientists and software engineers to deliver a spatial data science platform, develop industry-leading AI models for satellite imagery, and build world-class Geo AI solutions!

This challenge is a unique opportunity for you to solve a real-world problem! It will provide you an insight into the kind of work people at Esri are doing at the Esri R&D Center-New Delhi.

The problem statement is designed to evaluate your:

Familiarity with machine learning and deep learning models and frameworks.
Data augmentation skills.
Skill at training accurate deep learning models.
Documenting your work using well written Python code and Jupyter notebooks.

During the challenge, you will be expected to train a deep learning model. You can choose any library or methodology of your choice to train your model and need to upload your predictions in the specified format. You will also be sharing your notebooks and scripts needed to train the model for verification of results. Based on the submissions we receive, the top 30 participants will be invited to an in-person interview with the Esri leadership team in New Delhi from 18^th – 21^st March 2019.

Eligibility:

Current 2019 Batch candidates + Any candidate with 1+ years of experience.
Candidates from any branch or degree.

Challenge format:

1 Machine learning problem statement based on predictive analytics, deep learning, or/and libraries mentioned in the JD.

OPEN POSITION

Data Scientist

Experience:	0-10 years
CTC:	Best In Industry
Job Location:	Esri R&D Center - New Delhi

Responsibilities:

Develop and train deep learning models for computer vision problems such as object detection, image classification, road detection, building footprint segmentation, and 3D point cloud segmentation.
Develop processes and tools to monitor model performance and accuracy.
Integrate ArcGIS with popular machine learning and deep learning libraries such as Scikit-learn, Tensorflow/Keras, and PyTorch/FastAI.
Design, test, release, and support AI capabilities in the ArcGIS platform to enhance overall product quality and applicability for supporting data science and deep learning workflows.
Author and maintain geospatial data science samples using ArcGIS and machine learning/deep learning libraries like Tensorflow and PyTorch.

Requirements:

Knowledge of Python.
Knowledge of practical machine learning experience, some of which is within established technical organizations.
Self-learner with coursework in and extensive knowledge of machine learning and deep learning.
Experience with Python machine learning and deep learning libraries such as Scikit-learn, Pandas, PyTorch/FastAI, or TensorFlow/Keras.
Understanding of machine learning as well as deep learning techniques and algorithms such as k-NN, Naive Bayes, SVM, Random Forests, CNNs, RNNs, LSTMs.
Ability to design and implement deep learning models for object detection, semantic and instance segmentation, GANs.
Experience in data visualization in Jupyter Notebooks using matplotlib and other libraries.
Experience with Hyperparameter-tuning and training models to a high level of accuracy.
Ability to perform data extraction, transformation, loading from multiple data sources and sinks.

Recommended qualification:

Experience applying deep learning to satellite or medical imagery.
Familiarity with ArcGIS suite of products and concepts of GIS, including working with ArcGIS API for Python.
Knowledge of deep learning for natural language processing, probabilistic programming, and reinforcement learning.
Experience with CUDA/GPU programming.
Experience with distributed training of deep learning models and big data machine learning using Spark ML.
Familiarity with one or more of the following: Hadoop HDFS, Spark, Accumulo, Presto, MongoDB, Elastic Search, Cassandra, HBase, R, Mahout, Pig, Hive, DC/OS, Kubernetes.

ABOUT COMPANY

Esri is the global market leader in GIS, helping customers get results since 1969. Esri was founded to help solve some of the world’s most difficult problems. We do so by s...

more

GUIDELINES

This is an individual-participant challenge.
After you have built your application, you must do the following:
- Submit the solution in .zip format
- A...

more

This is an individual-participant challenge.
After you have built your application, you must do the following:
- Submit the solution in .zip format
- At the end of the page, you must complete the following tasks:
  - Upload the following as a zip, tar, or tar.zip archive:
    - Prediction file
    - Source file
  - Submit your solution
The submissions are evaluated automatically. The evaluation parameters include the following:
- Functionality of the code
- Design aesthetics
- Usability of the application
The evaluation of submissions can take up to a week's time.
When the challenge is live, your output will be evaluated only for 50% of the test data. After the challenge is over, your output for the remaining 50% of the test data will be evaluated and the final rank will be awarded.
If you do not select a submission file for the offline evaluation, your best submission will be automatically considered.
You will have to upload your submissions on the Problems page in the format specified in the problem statement.
In addition to your final submission, you will also have to submit your source file and other files as a .zip or .tar compressed archive.
The total number of submissions allowed for a participant throughout a challenge is 100. The maximum number of submissions that a participant can make in a day is 10.
You can use any tools or libraries to build your solution. There is no restriction on the tools that you can use.
Use of external dataset is prohibited for this challenge. Participants found using external dataset will be disqualified. A pre-trained model on a public dataset such as ImageNet may be used and fine-tuned.
Privately sharing code or data is not permitted—any case of code plagiarism will result in disqualification from the contest of all users involved.
Winners may be asked to demonstrate their work and share their codes.

By participating in this challenge, you are agreeing to HackerEarth's terms and conditions.

FAQs

Sample challenge

1. Can I participate in a sample challenge?

Yes, we recommend that you participate in our sample challenge.This challenge enables you to understand how to ...

more

Sample challenge

1. Can I participate in a sample challenge?

Yes, we recommend that you participate in our sample challenge.This challenge enables you to understand how to participate in Machine Learning (ML) challenges on HackerEarth. It comprises detailed information about how questions under ML challenges should be attempted and details about the HackerEarth judge.

Machine Learning challenge

1. Where will the challenge take place?

This challenge is an online challenge.

2. Will I receive a reminder notification before the challenge begins?

You will receive a reminder email 3 hours before the challenge begins only if you have registered for the challenge

3. How do I participate in this challenge?

To participate in this challenge, do the following:

Register for the challenge
Visit the challenge page on start date/time
Click Start now

4. How does the time limit work?

We follow an open time-limit format. In this format, you will have the following:

Challenge duration
- Duration between the start and end time of the challenge. For example, 24 hours (6 p.m. on Wednesday up to 6 p.m. on Thursday)
Challenge-completion duration
- Time within which you must complete the challenge. For example, within 12 hours (starting any time after 6 p.m. on Wednesday)
- After you start the challenge, you cannot pause it

5. The challenge is open for 24 hours (6 PM on Wednesday up to 6 PM on Thursday). If I start the challenge on 5 PM on Thursday, how long will I get to complete the challenge?

This is a fixed event, which means that the duration is fixed. The challenge will end on the date and time that is given on the challenge page. The amount of time you will get to complete the challenge is based on when you begin the challenge.

If you register for and start the challenge on Thursday at 5 PM, you will get only 1 hour to complete the challenge.

6. How are the Machine Learning problems evaluated?

The evaluation is based on the evaluation metrics that are defined for a problem. When the challenge is live, your submission will be evaluated only for 50% of the test data.

After the contest is over, we re-evaluate and assign a score to your final submission based on the complete data set. Therefore, the score after the offline evaluation is subject to change.

7. How can I submit my final submission file?

You can submit your output by uploading your submission file on the Problems page in the format specified in the problem statement.

8. What all should I submit other than the submission file?

In addition to your submission file, you must submit the following:

Text file explaining your approach to the problems
Source files used to build your ML models

All these files should be submitted as a compressed archive in any one of these formats:

zip
tar
zig

9. Which tools can I use for building my models?

You can use any tools or libraries for building your models. There is no restriction on thetools that you can use.

10. Where will the challenge take place?

This challenge is an online challenge.

11. Can I participate in a challenge on a mobile platform?

Currently, HackerEarth does not support mobile platforms for participating in any challenge. We recommend that you should participate in a challenge using a laptop or desktop.

12. Is the online phase of a ML challenge restricted to any specific programming language?

The online phase of the challenge is not restricted to any specific programming language. You can code in any language you prefer but the submission file should be submitted in the specified format.

13. What will happen in case if there is an issue with my computer or Internet connection?

If there is an issue with your computer or your Internet connection, you will not lose your previously submitted answers. They will be saved on HackerEarth's servers. You can log in again and continue coding from where you stopped.

However, the timer will continue running, which is why it is advised that you ensure that your computer/Internet connection is working properly before you begin the challenge.

14. Whom should I contact in case of submission or any other failures?

If there is an issue related to your submissions, please send an email to support@hackerearth.com.

15. Who do I contact if I face an issue during a live challenge?

For any queries/issues/support, you can contact us in the following ways:

Live IRC (Freenode network) at #hackerearth
Send an email to support@hackerearth.com

16. How do I submit the prediction file?

To submit your submission file, follow these steps:

Navigate to the Problems tab.
In the Upload Prediction File section, upload your prediction file in the format specified in the problem statement.
Click Choose File.
In your system, navigate to the respective folder where you have saved your prediction file.
Click Open. You will be redirected to the Problems page.
Click Submit & Evaluate to evaluate your prediction file.

You have successfully submitted your prediction file.

17. How do I submit the source files?

To submit your source files, follow these steps:

Navigate to the Problems tab.
In the Upload Source File section, upload your prediction file in the format specified in the problem statement.
Click Choose File.
In your system, navigate to the respective folder where you have saved your source file.
Click Open. You will be redirected to the Problems page.
Click Upload to submit your source file.

You have successfully submitted your source files.

18. Is it possible to change my submission after I have submitted a file?

Yes, you can edit and submit the solution to a question (that you can view and edit) any number of times before the challenge-completion window closes.

Note: Some tests have timed sections. After the time limit has been reached, you will not be able to revisit questions in that section.

19. Where can I view my submissions?

You can view your submissions in the Submissions section of the Challenge page.

Esri Data Science Challenge 2019

Social Share

ANNOUNCEMENTS

ABOUT CHALLENGE

OPEN POSITION

ABOUT COMPANY

GUIDELINES

FAQs

Sample challenge

Sample challenge

Machine Learning challenge