Currently working as a Deep Learning Engineer at Intel in Healthcare AI.
Technical mentor for healthcare based startups under plugin accelerator program.
Working with Movidius Neural Computing Stick.
Previously contributed to projects related to connected homes, smart city, mobility and edge AI.
Exhibited the contributions to NCS and won best exhibit award for Intel at VLSID '18 and ES '18 conference.
Represented Intel at NASSCOM India Leadership Forum '18 & World Congress on Information Technology '18; IEEE International Conference on Electronics, Computing and Communication Technologies '18; Open Data Science Conference '18; Grace Hopper Celebration India '18; India Innovation Showcase; IEEE International Conference on High Performance Computing, Data and Analytics.
Initiative to make research accessible. Worked on the most influential papers from top conferences in AI. The role of DRI further includes the task of leading and managing the group along with contributing.
Shipped features to introduce personalisation and target push notifications to increase CTR & reduce spam/ unsubscriptions. The task involved working with basic probability and statistics, various clustering and classification models, PyLDAvis, Sillhoutte analysis, basic data cleaning. Improved the skills on scalability, code conventions, pair programming.
Ideated and worked on implementing analytics and machine learning algorithm to the R&D CRM product being developed.
Built a model to predict machinery failures using bearing dataset. Applied deep learning on sound for BPO use case. Further, worked on Bosch dataset as per the requirement.
My team ideated a business simulation product and later developed a gamified version of the simulation product using Unity 3D. GUI was inspired from Sim City and back-end logic was framed as per the company's requirement for the product.
Refracted the code base of processing.py and made the entire code base of sample examples error free.
Will be speaking on Rethinking privacy in the age of AI using differential privacy for deep learning applications.
Spoke on the topic of Artificial Intelligence on the Edge.
Demonstrated the power of the Neural Compute Stick and OpenVINO toolkit for developing affordable AI at the network edge.
Published an algorithm to improve the performance of PageRank algorithm in terms of time complexity.
Best project award for breast cancer detection and the most popular award for stock price movement prediction.
Academic rank of top 4/219 throughout the degree.
Exhibited the contributions made to the Neural Compute Stick.
Awarded for being Results Orientation. The value tags given in the recognition included- Execute flawlessly, set challenging and competitive goals, constructively confront and solve problems, assume responsibility, focus on output.
The project was related to the prediction of flight delay.
Awarded for student achievement and contributions.
During the hackathon, created a user-friendly server management console panel.
Selected and invited to MIT Media Labs emerging world program, JP Morgan code for good, 1st ever Google Developer Days, India.
C, C++, Python, R, MATLAB, Web development
sklearn, TensorFlow, numpy, OpenCV, etc
Machine learning, Deep learning, Edge AI, Data analytics, NLP, Graphs
Modified Dijkstras algorithm and ant colony optimisation to arrive at the destination in the shortest time possible. Working on integration with deep learning techniques in an attempt to create a complete automated transportation model.
Developed a multi-class audio classifier on large dataset (>20GB) and further working on audio content analysis.
Deep learning model to detect abnormalities in the brain using MRI and CT scans.
ML algorithm to predict vehicle trajectory and the estimation of fuel consumption and CO2 emission by the vehicles.
A simulated 3D guidance system starting from the entrance of the airport to the aircraft using Unity.
Working on the development of a device to help the visually impaired to read using edge AI and bone conduction.
Variation of Conways game of life including gender and graph for better accuracy in prediction suited to human beings.
Automated calculation of centrality measures for large graphs.
Automatic caption removal from images using OpenCV.
Sensor node to detect forest fires and transmit the information to the concerned authorities.
Model to simulate the growth of forest from scratch using cellular automata and WebGL.