Abhinav Ramesh Email & Phone Number

Chennai, Tn, In

Bengaluru, Karnataka, India

Bengaluru, Karnataka, India

London, England, United Kingdom

Geneva, Switzerland

• Devised an efficient scheme involving denominating bins for allocating memory blocks and caching them for reuse in order to minimize allocations and increase event processing throughput for pixel tracks and vertices reconstruction in the CMS detector within the LHC• Implemented allocators for host and device memory management based on this scheme using C++ along with Alpaka(platform-independent parallel programming library), and exposed this functionality via a smart pointer interface, resulting in a speedup of 2.5 for CUDA

Edinburgh, Scotland, United Kingdom

• Engineered algorithms to detect T-junctions and associated priority roads, and to order the onward side-lanes of a particular side-lane from left to right, given a road network map, for determining right of way, assisting human annotators and helping agents in simulations• Implemented these algorithms using Python, and incorporated them into a visualization tool as sub-tools that can highlight t-junctions along with priority roads as well as indicate turn directions for a route on a map

Tirupati, Andhra Pradesh, India

• Developed graph algorithms for finding inter-town journeys by optimizing multiple criteria(arrival time and transfers) and considering multiple modes of transport, by splitting the computation into preprocessing and run-time phases• Employed Java to generate and store optimal station transfer pattern graphs in Neo4j, used Ruby to store direct connections between stations in SQLite, as preprocessing, and wrote a PHP web app to invoke Java at run-time to compute optimal journeys by querying these databases

Manama, Capital Governorate, Bahrain

• Created dashboards in IBM Watson and Cognos Analytics, utilized Numpy/Pandas for data analysis and generated interactive visualizations using Bokeh for H.R. team to transform metrics into strategic decisions• Constructed a classification model using Scikit-learn to accept or reject loan requests and achieved an ROC-AUC score of 0.96

Master Of Science - Ms, Computer Science, 4.00/4.00

Activities and Societies: Phi Kappa PhiCoursework: Advanced Algorithms, Machine Learning, Parallel Systems, Reinforcement Learning.

Bachelor Of Technology - Btech, Computer Science, 9.46/10.00