Meet The Winners of Latest MachineHack Hackathon: Work Hour Prediction Challenge

The challenge was a part of the new MachineHack Fortnight Hackathon Series, where unique problem statements are provided every week to test one's data science skills and abilities.


The Fortnight Hackathon Series for Data Scientists based on the Work Hour Prediction Challenge was successfully conducted and concluded on 1 October 2021. The challenge was a part of the new MachineHack Fortnight Hackathon Series, where unique problem statements are provided every week to test one’s data science skills and abilities. The main challenge of the hackathon involved predicting the working hours per week at different locations with attributes such as workclass, education, marital-status, occupation capital-gain, capital-gain, capital-loss, etc., to get the desired salary in a range. 

The relevant skills that were required to solve such a challenge were:


Sign up for your weekly dose of what's up in emerging technology.
  • Optimize RMSE
  • Forecasting
  • Timeseries
  • Machine Learning Approach

The submission limit for the hackathon was set to 3 per day; the Work Hour Prediction Challenge was initiated on 17 September 2021 and received a huge amount of positive participation.

Rank 3 – Nitin Rai

After his graduation, Nitin joined an MNC to work as a Systems Engineer. He later started reading and learning about data science online and began exploring it. He also did a course on Big Data Analytics and took up a job as a Data Scientist in a startup, which he left later to take up a job in a different domain.

Approach Used

Carefully understanding the problem statement and data provided, Nitin understood that the problem is a regression type. He used the CatBoost algorithm, which comes with gradient boosted decision trees. In order to avoid any overfitting, he decided to use cross-validation techniques. Nitin made use of 5 fold validation and built five CatBoost models, and predicted five values. The weights for ensembling the predictions were obtained from the local scores. He also inversed the scores, as a lesser RMSE score meant better fit, and used them as weights for ensembling. 

Nitin’s MachineHack Experience

Nitin said that “Hackathons like these boost the confidence of any aspiring Data Scientist. I had a great time participating in the Hackathon and would like to participate in any hackathon hosted in the future as well”.

Nitin’s solution can be viewed by clicking on the link here

Rank 5 – K V Shivaramakrishna

KV is a Data Scientist at Gramener, with 5+ years of experience in the field of Machine Learning and with a special interest in the field of Computer Vision and Natural Language Processing. He has a background in Electrical Engineering with a Masters in Electrical Machines and Drives. KV joined IBM right after college into the Cognitive Practice to develop ML/AI-based applications for customers in multiple industries like retail, BFSI, Pharma, etc. 

Approach Used

KV performed EDA to check the quality of data, output labels and other features. He used OneHotEncoding to convert the categorical columns. Trying multiple models like CatBoost, LightGBM, XGBoost, etc., on the pre-processed data with default hyperparameters, he tried fine-tuning all the models using Optuna for hyperparameter tuning. Doing so, he got the model with the lowest RMSE using the XGBoost and tuned hyperparameters. 

KV’s MachineHack Experience

KV said, “The participation helped me judge my own understanding of the basic ML/Dl techniques. MachineHack is a great platform for young enthusiasts to hone their skills and be ready for interviews and other hackathons”.

KV’s solution can be viewed by clicking on the link here

Rank 6 – Ayush Patel

Ayush’s journey began when COVID-19 lockdown was imposed, where he came across Krish Naik’s numerous video tutorials on Data Science which fascinated him, and he started learning Python. He created a mini project on Face Recognition System and a Crop Disease Detection app during his final year pursuing BCA. 

Approach Used

Ayush tried Label Encoding and One Hot Encoding for Categorical features, but nothing improved much. For his final approach, he decided to replace categories with the mean of the working hours for a particular category and check the score with both linear models and tree-based models. He also chose the best model and then passed it to the Voting Regressor.

Ayush’s MachineHack Experience

Ayush said, “I think MachineHack is a great platform for participating in hackathons. I personally prefer MachineHack because I am still at rank 1 in book price prediction; it’s the first challenge I won, so whenever I feel demotivated, I always take a look at the leaderboard”.

Ayush’s solution can be viewed by clicking on the link here.

Heartiest Congratulations to the winners of the Work Hour Prediction Challenge of MachineHack Fortnight Hackathon Series Edition – 1.

More Great AIM Stories

Victor Dey
Victor is an aspiring Data Scientist & is a Master of Science in Data Science & Big Data Analytics. He is a Researcher, a Data Science Influencer and also an Ex-University Football Player. A keen learner of new developments in Data Science and Artificial Intelligence, he is committed to growing the Data Science community.

Our Upcoming Events

Conference, in-person (Bangalore)
MachineCon 2022
24th Jun

Conference, Virtual
Deep Learning DevCon 2022
30th Jul

Conference, in-person (Bangalore)
Cypher 2022
21-23rd Sep

3 Ways to Join our Community

Discord Server

Stay Connected with a larger ecosystem of data science and ML Professionals

Telegram Channel

Discover special offers, top stories, upcoming events, and more.

Subscribe to our newsletter

Get the latest updates from AIM