Senior Algorithm Engineer
Intelligent Incident Management System
Lead the team designed a data-driven algorithm for router alarm clustering, incident tracking & root cause analyzing system with high flexibility which can adapt in different network environments
Handbook Knowledge Extraction
Algorithm Engineer Intern
GPS coordinate correction system
Using Scala, based on OpenStreetMap (global road map) and vehicle GPS equipment data, a design based onTime-series GPS correction system with Viterbi algorithm and real-time update to Hive database
M.S. in Electrical and Computer Engineering
2020.01 - 2020.02
Disigned a system able to parse and execute programs written in the assembly language, and modeling the behavior of the programs as if they were executing on a simple computer processor. The system also involved memory allocations and extensive Junit tests.
2020.01 - 2020.02
Constructed a cryptarithm solver system. The system includes context parser, an efficient general-purpose permutation generator using Heap's algorithm, a context expression converter, a calculation operator and a solver to get all the results to decode the cryptarithm.
AOD-DC Image Dehaze Network
2019.09 - 2019.12
Proposed an image dehaze system by combining a convolutional neural network (CNN) with dark channel prior named AOD-DC. Achieved a good visualized dehazing result from blurry images using Google Street View and NYU Depth V2 dataset.
End-to-End speech to text transcription System
Designed an end-to-end system for speech to text transcription with combination of techniques of LSTM, CNN, Encoder, Decododer and Attention method for WSJ dataset, the final result achieved less than 10 Levenshtein distance on test dataset after 60 epochs.
Research Assistant at Chongqing Big Data Lab
Movie Recommender System
Participated in the research project on the recommendation system, and proposed a high-dimensional sparse algorithm based on the use of elastic network and latent feature model According to the recommendation algorithm above, the obtained results are improved by 0.001RMSE compared with the traditional latent feature-based model.
EEG Signal Processing Survey
Participated in the research on the extraction and classification of EEG signal features in recent years, and summarized related algorithms