• Conducted AI research primarily in Natural Language Processing (NLP) and Computer Vision (CV). Specific research area including but not limited to sentiment analysis, question generation, automatic medical report generation.• Organized internal and external AI research collaboration.Achievements:• Best paper award at 2024 IEEE Conference on Data and Software Engineering (ICoDSE)

• Summarized 800+ village demographic attributes from 84k+ Indonesian villages in a Tableau dashboard.• Performed data cleaning and wrangling.• Conducted sentiment analysis of some public figures (continuing the intern project).• Empowered the front-end developer in creating the web UI using Nuxt.js.

Entrusted two main projects:1. Media Monitoring• Executed sentiment analysis research on Indonesian figures having a potential to be the next Indonesian president in 2024.• Automated the process of collecting text data from websites and social media through scheduled web scraping.• Implemented machine learning and deep learning models for sentiment analysis. Diagnosed the cause of model deficiency. Experimented with some feature engineering and other techniques to handle it.• Created a dashboard to visualize the media monitoring results.2. Potensi Desa• Reviewed the requirements for tourism village attributes. Collected the necessary data based on requirements analysis through web scraping.• Utilized big data tools for data cleansing, data exploration, and data wrangling.• Learned the necessary tools to build map visualization for further project development.

Tutored : • Basic programming using Python (Python data types, if-else in Python, Python loops, Python data structures)• Advanced programming using Python (Python functions, high order functions, class and modules)• Search algorithm (depth first search, breadth first search, Dijkstra algorithm, greedy search, A* search, constraint satisfaction problem) concepts and implementations using aima-python module. • Proportional logic, first order logic and inference (DPLL, forward chaining and backward chaining) concepts and implementations using aima-python module.• Naive Bayes Classifier, Bayesian Network and Decision Tree concepts from scratch.

The project is a year temperature forecasting (from September 2022 - September 2023) in Jonggol District, Bogor Regency to determine the optimal time to plant sorghum in Jonggol. This project is done as a completion task of GEMASTIK XV Data Mining Competition elimination round.• Reviewed 13 literature related with land suitability, food security, sorghum, and LSTM model to discover solutions for the food security problem forecasted by current Indonesian vice president.• Collected ten years of Jonggol temperature data using QGIS and Visual Crossing API, analyzed it to determine the LSTM model forecast period, then preprocessed it to satisfy LSTM input using Pandas and NumPy. • Engineered an LSTM model using Tensorflow to forecast temperature in Jonggol district to determine its land suitability to plant sorghum in one year period. The model has an MSE value of 0.5.• Created a calendar to help farmers in Jonggol district determine the optimal time to plant sorghum in the period from September 2022 - September 2023 based on the built LSTM model forecast.

The project is a sentiment analysis of Indonesia farm land use change from 2017 - 2021 for Indonesia GEMASTIK XIV national competition held by Ministry of Education in 2021.• Conducted sentiment analysis research on Indonesia farm land use change issue by collecting tweets from 2017 until 2021 related to the issue from Twitter using Twint API.• Annotated the tweets using InSet for classification purposes, cleaned them using Python regex, then preprocessed the cleaned tweets by transforming them into TF-IDF vector to satisfy the machine learning models input.• Built a logistic regression and a SVM model using Scikit-learn to classify positive or negative sentiments, then optimized the hyperparameters using Optuna. Logistic regression as the chosen model has an accuracy of 0.75.

The goal of the hackathon project is to build a classifier to classify whether an employee can be categorized as Best Employee or not.• Visualized the dataset using Matplotlib and Seaborn.• Preprocessed the data by normalizing and encoding the features in the dataset using Pandas.• Built Adaboost, Random Forest and MLP model and assessed the goodness of the model using a learning curve using Scikit-learn.• Fine-tuned the built model using Optuna.