- Engineering manager and tech leader for the Knowledge Engineering System team.- Led the core project WishBlue Factory-Direct, which, being the single source of truth for product listing, supplier profile, inventory info, backs multiple projects (1sansome.com, wish.com, etc.);- Led the Managed Merchant Service project, to fully support the launch and operation of Wish Fashion successfully, in a very limited amount of time.- Led the team on designing the architecture and solutions for the projects and some of its related cooperative projects;- Implemented the frameworks and business flow for the project, from the web API server to backend services. Major work is in Golang, some in JavaScript and Java.- Led the direct communications with different business teams, to provide solutions, priorotize features, and plan engineering resource.

- Worked on enriching the visions, and instanciate them into executable product blueprint.- Built up a highly capable engineering team from scratch in two month; Designed the organization, architecture, procedure, and culture.- Managed the iteration of both product and engineering work, and shaped the cooperative routine to improve efficiency.

- Lead the team to design and implementation of high performance event driven micro-service backend architectures.- Implement internally standard data contract libraries for efficient tenant aware object storage and query management across No-sql databases(Couchbase, Firestore) and Kafka.- Lead the data engineering team to implement data pipelines for streaming million level daily data from webhook to Kafka, and eventually to BigQuery. And build tools to make them easily usable.- Design and implement AI methods (e.g. Interest Graph) to mine the data and generate knowledge for targeting and personalization.

Founded Urec.io (Non-profit, originally as Geekrec) to be the first mover in the recruiter business review and scoring industry.Recruited top tier talents and built a team of 12, including business (branding, marketing, operation) specialist, content creators, UI/UX designers, IT engineers.Led the team to drive the project from idea, all the way through designs, demos, and eventually the version 1 product in less than 6 months.With this unpaid volunteering team, no milestone was missed, the first launch was full of success, and all initial requirements was met.Designed and built the system for https://urec.io and https://blog.urec.io from scratch as the interim CTO at the very beginning, before we have an offical CTO.

Working in the machine learning group, Snapdragon Neural Processing Engine (SNPE) team.- Worked on SNPE project;- Integrated lower level components;- Compiled solutions for couple major issues;- Working on Android Neural Network (ANN) API project; Implement asynchronous neural network computational frameworks for ANN, on CPU, GPU, DSP, etc.- Cooperating with Google ANN team on feature design, interface implementation, and optimization.- Developing neural network troubleshooting tools.

- Worked in J9 Runtimes team.- Re-built the existing automation tooling software in Python 2.7, in order to ensure the platform-independency for these programs. Also isolated the platform dependent code in lower level libraries, to fit different platforms like x86, ppc*, etc. - Transformed the old production system from closed IBM-owned software solution to white water solutions, like some open sourced projects. I had to build adapters, or converters, to make sure these transformations go smoothly, and re-represent as much old data as possible.- Built Python-driven RESTful APIs for data transmitting, database query, etc. Because the system consists of subsystems located on different cities, the old way of data transmission and communication was not secure and decoupled. So I tried to hide each of the server site behind its own http server, and standardize a set of RESTful APIs for inter-site communication.- Re-architected the build automation system, so that the build time remarkably decreased by 90%, and the stability, efficiency, flexibility all got improved. This gained a high reputation for our team.

Main responsibility: Machine Learning and NLP R&D- Designed and developed a large-scaled Tweet stream summarization system, based on Topic Modeling, Machine Learning, webpage scraping, etc. 1) The system uses optimized TF-IDF based algorithm to filter a set of meaningful terms from clusters of tweets, and apply them as corpus to train a Latent Semantic Analysis (LSA) model continuously. The LSA model is then used to evaluate a cluster of tweets, and represent the most essential topics.2) For evaluating topics from webpages in posted links, I developed and deployed a distributed webpage scraping program, which use Spark as the distributed computing platform, and add the results to optimized the topic summarization of Twitter text.3) After generating the feature vectors for representing each cluster of documents, the system also support evaluating similarity of a given document against each of the clusters. - Designed and developed a scalable real-time Twitter event and trend monitoring system. This system is used to help customer to real-timely monitor the events happening on Twitter that are related to them, like new follower, hashtag, @, direct message, etc. These collected information will then be processed and analyzed periodically to generate reports, ready for front end to visualize. The system utilizes a lot feature of AWS, like real-time stream process model and toolkit, cloud based Non-rational database, automatic load balancer etc.- Took part in the system administration, as a member of our compact team. Since every developer was working on fairly different parts, I had to built and maintain most of the infrastructure and dependency that was directly used and related to my projects. These works are all done on *nix system.

- Worked in a team for a industrial Super-Resolution (SR) project with Alentic Microscience Inc.- Focused on SR reconstruction algorithms and analysis of medical images, - Additionally, implemented image preprocessing, efficiency analysis, algorithm implementation, etc.- Programmed in Matlab to design and implement experiments.

- Set up from scratch, being responsible for most of investigations, planning, business communications, etc.- Managed a team with 4 members and up to 5 staffs, doing Advertisement, marketing, and data analysis by myself.- The corporation grows very well until now.

- Worked for one top team in the medical physics field in the world, on Image-Guided Radiation Therapy.- Focused on deformable 3-D CT image registration, analysis, data visualization, etc., using ITK and VTK, in C++.- Highly praised for my research ability, by the supervisor and colleagues. (Reference available upon request)Publications:- Hybrid deformable registration method joint using TPS and B-spline in CT image, International Conference on Biomedical and Health Informatics BHI-2012, IEEE-EMBS- Deformable Registration Method using TPS and B-Spline, Image and Signal Processing in Health Care and Technology (ISPHT 2012) HCI, CIIT

- Concentrated on image processing, fingerprint feature extraction, etc.- Contributed to a granted patent on a fingerprint identification system used on automobile, as a core member in the team.- Programed in Matlab for analysis; and C language, assembly language for hardware implementation.

- Led to found this first technical society in Chongqing University.- The most influencing student organization with over 200 members.- Managed software development project efficiently, and organized public educations.- Cooperated with Security Industries, such as Kaspersky and ETS.