PRAKAS SHRESTHA

PRAKAS SHRESTHA

Data Analyst & ML Engineer | Python · SQL · GenAI

Data Analyst and Machine Learning Engineer with hands-on experience in end-to-end data pipelines — from raw semiconductor data cleaning to deploying ML models.Skilled in Python, PostgreSQL, Snowflake, Deep Learning, NLP, and Generative AI (LLMs, AI Agents, Prompt Engineering).


🛠 Skills

Python Pandas NumPy Matplotlib Seaborn Scikit-learn Linear Regression Logistic Regression Decision Tree Random Forest XGBoost K-Means Clustering SVM Deep Learning Neural Networks TensorFlow Keras NLP Transformers LLM Generative AI Prompt Engineering AI Agents LangChain Data Cleaning EDA Feature Engineering PostgreSQL Snowflake SQL ETL Semiconductor Data Analysis

💼 Experience

Data Scientst

🏢 Sumitomo Electric

Feb 2023 — Present

Semiconductor Manufacturing Defect Analysis Project

This project aimed to identify root causes of chip defects in semiconductor manufacturing and reduce defect rates. We collected and integrated process data (temperature, pressure, etc.) and equipment operation logs, managed data centrally, and conducted analysis to identify defect causes.

Key Responsibilities:

Data Collection & Preprocessing (ETL Processing) Automated collection and preprocessing of raw data including device logs, sensor data, and production records using Python and SQL. Extracted data periodically from multiple sources (equipment CSV outputs, MES systems, external databases). Performed data cleaning including removal of anomalies, missing value handling, time-series alignment, and field normalization. Integrated processed data into Snowflake data warehouse to build high-quality data foundation for manufacturing analysis, defect cause analysis, and AI model training.

Analysis using Sonar Analytics Tool Conducted preprocessing, visualization, and analysis of manufacturing data (device logs, sensor data, quality results) on Sonar BI/ML platform. Performed correlation analysis, trend analysis, and anomaly detection to identify critical variables and equipment status. Built anomaly detection models using Sonar's machine learning capabilities. Implemented automated anomaly alerts coordinated with Statistical Process Control (SPC) for real-time monitoring and predictive maintenance.

Data Scientst

🏢 Metic Fielder

Apr 2019 — Jan 2023

Wearable Biosensor-based Worker Health Management System

Developed wearable sensors and smartphone application for monitoring worker health conditions. The system estimates body temperature and physical load from biometric data and predicts heat-related illnesses in real-time, sending alerts to workers via smartphone app while storing and analyzing data on the server.

Key Responsibilities:

Smartphone App Development Developed Android application in Java with body temperature estimation logic. Receives biometric and environmental data, calculates health metrics, and sends real-time alert notifications to workers.

Wearable Data Analysis Analyzed wearable sensor data collected from workers for health prediction. Created statistical analysis programs in Python and automated data processing using VBA macros.


🎓 Education

大学

Apr 2015 — Mar 2019

🏫 第一工科大学東京上野キャンパス

工学部


🏆 Certificates

🏆

No certificates yet!


🔥 Recent Projects

🌟

No projects yet!


✍️ Latest Posts

✍️

No blog posts yet!

We will be posting soon. Stay tuned!


📈 GitHub Stats

GitHub Stats
Top Languages
GitHub Streak