Software Engineer - WSO2

            ( Feb 2025 - May 2025 | 4 months )

Data Science Engineer - Air Arabia Airlines (ISA)

            ( Feb 2024 - Jan 2025 | 1 year )

Responsibilities

1. Flight Delay Prediction: Predicting departure and arrival delays of flights for next 24 hours, so that airlines team could swap the aircraft for different flight assignments.
   • Data Cleaning and Feature engineering (e.g., Moving average for different aggregations such as departure delay specific to the origination airport)
   • Trained separate models for arrival and departure delays and experimented with different algorithms: XGBoost, RandomForest, LightGBM
   • Automated hyperparameter tuning with Random Search and Bayesian search
   • Cross Validation (Time series split) and calculation of Feature Importance
   • Evaluated model performance using MSE, RMSE for different stages in the sequence
   • Experimented with Error modeling to reduce error propagation
   • Deployed training and prediction services with Docker and Jenkins
   • Development of Monitoring service and its deployment
2. Flight Fuel Burn Prediction: Airlines want to assign best fuel efficient aircrafts to the longer journeys. So that they can reduce cost
   • Data Cleaning and Feature engineering (e.g., Historical Average for different levels of aggregations: Historical averge for different pax level, since pax level is the high importance feature)
   • Experimented with different models: XGBoost, LightGBM, CatBoost, MLP
   • Experimented with ensemble approach (stacking approach)
   • Meta model training using only predictions of base models using Bayesian ridge, linear regression, XGBoost, and decision tree
   • Meta model training with both features and predictions using XGBoost
   • Experimented with stabilizing the variance of the target variable using log, sqrt, and box-cox methods
   • Evaluation of the model using MAE and MAPE

Trainee Software Engineer - 99X

            ( Dec 2022 - June 2023 | 6 months )

Responsibilities

• Migrated system API Gateway to Spring Cloud Gateway from Spring Zuul
• Implemented Liquibase Context for environment restriction
• Worked with Cucumber for BDD Testing
• Developed Docker in Docker (DinD) solutions
• Implementation of Execution Context in Spring Batch
• Creating a Dockerfile that runs a JMeter script and generates an HTML report
• Some R&D work to incorporate new ML models into the system