Machine Learning Information Technology

Rohan Kuckian

Stony Brook, NY *****.*******@*****.*** (934) 255 – 9006 LinkedIn Github

EDUCATION

Stony Brook University Stony Brook, New York

Masters of Science in Computer Science (GPA – 4.0/4.0) Aug 2024 – May 2026 Relevant Coursework: Data Science Fundamentals, Analysis of Algorithms, Human Computer Interaction University of Mumbai Mumbai, India

Bachelor of Engineering in Information Technology (GPA – 9.78/10) Aug 2018 – Jun 2022 Relevant Coursework: Database Management, Machine Learning, Web Development, Software Engineering TECHNICAL SKILLS

Languages: Python, SQL, Java, HTML, CSS, JavaScript, Node.js Frameworks: PyTorch, NumPy, Pandas, Matplotlib, HuggingFace Databases: Oracle 19c, MySQL, MongoDB

Developer Tools: Git, Jupyter Notebook, VS Code, Eclipse, JDeveloper, SQL Developer, WebLogic Server, Jira Data Analysis Tools: PowerBI

EXPERIENCE

Oracle Financial Services Software (OFSS) Mumbai, India Associate Consultant Jun 2022 – Jun 2024

● Architected and customized critical banking interfaces and business logic for Flexcube's Universal Banking and Payment systems, handling 20M+ daily transactions.

● Engineered and developed automated PL/SQL package solutions to process 1M+ Calypso-generated transactions, significantly reducing manual intervention and operational overhead.

● Spearheaded implementation of services using Oracle Banking Microservices Architecture (OBMA), enhancing system scalability and overall product performance.

● Designed and deployed cross-platform APIs to enable seamless integration between three core banking applications, achieving 4x improvement in API delivery efficiency.

● Solved 10-20 critical issues arising in UAT and Production on a day-to-day basis, ensuring smooth handover of system to end users.

● Earned the prestigious Dashing Debut Award in recognition of outstanding first-year performance and technical contributions.

PROJECTS

Distributed File Storage System Python, gRPC Aug 2024 – Dec 2024

● Designed and built a distributed file system using Python and gRPC that supported scalable file storage and retrieval across 5+ nodes, achieving 99.9% availability under simulated network failures.

● Implemented Raft-based leader election and heartbeat monitoring, reducing failover recovery time to under 3 seconds and ensuring consistent performance across 100+ simulated client requests per second.

● Optimized data chunking, replication, and load balancing strategies, leading to a 60% improvement in upload/download throughput and a 40% reduction in latency during parallel file access. LLM Performance Optimization Python, PyTorch, Hugging Face Jan 2025 – Present

● Developed performance tests for various LLM architectures, analyzing model inference speeds.

● Implemented quantization and k-v caching optimization techniques, resulting in a 33% improvement in model inference time.

Predicting Business Domain of Companies with their Logos Python, PyTorch, Pandas Nov 2024 – Present

● Prepared and executed web scraping algorithms to harvest 100,000+ corporate logos from search engines, creating a comprehensive dataset.

● Used DenseNet architecture for outlier detection and leveraged an ensemble of ResNet models to achieve 25% accuracy in logo classification.

● Incorporated OCR technology for text extraction from logos, resulting in 41% accuracy for business domain categorization across 8 industries.

Contact this candidate