🚀 My Projects
🔍 Beneath The Surface: Scene Perception with Depth Estimation
📌 Tech Stack: Python, ResNet-50, MTI-Net
🛠 Developed a multi-task learning framework using MTI-Net to unify depth estimation, semantic segmentation, and edge detection, enhancing indoor scene understanding.
📊 Achieved depth loss of 0.0339 with advanced guided attention mechanisms and multi-scale feature fusion for superior depth processing.
🔗 Click To Read the Research Writeup
📰 Tell IF Fake: Fake News Detection on Twitter
📌 Tech Stack: Python, TF-IDF, LDA, GMM, RoBERTa, Naïve Bayes
🛠 Implemented TF-IDF weighting, Latent Dirichlet Allocation (LDA), and Gaussian Mixture Models (GMM) to improve text analysis in fake news detection.
📊 Integrated pre-trained RoBERTa for sentiment analysis, achieving high accuracy in differentiating fake vs. real news.
🔗 Click To Read the Research Writeup
🚗 Seat Share: Mobile Ride-Sharing App
📌 Tech Stack: React Native, Firebase, PostgreSQL, Stripe API, Redis
📱 Building a React Native mobile app for seamless ride-sharing on iOS & Android, integrating real-time updates via Firebase and secure payments with Stripe.
⚡ Optimized performance using Redis caching and PostgreSQL for scalable backend operations.
🏢 Industry Work
📈 Observability at Getir
📌 Tech Stack: AWS Redshift, New Relic, Slack API
🚀 Developed New Relic dashboards for real-time monitoring, identifying critical bottlenecks and improving response times by 20x.
🔔 Built an automated Slack alert system for proactive error detection, ensuring stable microservices and faster, more informed critical incident response.
📊 Portal for Uniper Trading Analytics
📌 Tech Stack: Flask, Python, Azure, Snowflake, Excel-Python Integration
🛠 Designed a Flask-based web application to aggregate & analyze trading data from multiple vendors using Azure and Snowflake.
⚡ Optimized system performance by implementing data caching, async task queues, and CI/CD pipelines with Spinnaker.
Want to see more? Visit my GitHub!