AI & ML
Models, experiments, and intelligent systems built for tasks like classification, generation, and optimization.
-
Predicting future video frames to enhance and interpolate low-frame-rate video streams. The task involved using LSTM architectures with squeeze-and-excitation modules to enhance performances.
-
First automatic translator and restorer for ancient Greek Linear B texts using multiple smaller models. Final translation is done by asking Gemini LLM to pick best reconstruction among options.
-
Research project on optimizing client selection in Federated Learning using Weighted Local Rademacher Complexity and other statistical learning tools.
-
Implemented Q-Learning, DQN, and DoubleDQN in Atari environments for comparison of performance and sample efficiency.
Full-Stack Development & DevOps
Web apps, APIs, cloud integrations, and deployment pipelines using modern tools and frameworks.
-
Blogging website about music built in PHP. Features include comments, likes, trending posts, and user ranking.
-
Ruby on Rails tourism platform with messaging, maps, Stripe integration, ratings, and personalized accounts.
-
Microservice-based AI feedback system with login, database, inference API, ELK stack, and monitoring tools.
-
This personal portfolio built with Jekyll, HTML/CSS, and fully responsive Apple-style UI.
Data Engineering
Data pipelines, transformation jobs, and analytics projects focused on performance and scalability.
-
Setup of a 3-node HBase cluster for distributed big data querying. Included a tutorial and benchmarking.
-
Built an ETL pipeline using Pentaho to reconcile highly heterogeneous datasets from multiple sources.
-
Performance and use-case comparison between Graph and Relational DBs using analytical queries.
-
Mini search engine for Amazon product listings built using PySpark with tf-idf scoring.