Designed a system processing social media data. Text classifier and image classifier is applied to social media content. Implemented a prototype application for iOS using Swift, Python and Core ML. Trained the text classifier with 70, 000 labeled social media data content. Developed techniques to evaluate the models.Transferred social media data with Rest API and Webhook (Google cloud and DialogFlow).Filed a non-provisional patent and a research paper regarding the social media content processing system.

Developed an application for iOS which gathers energy data on IoT devices and uses data visualization to analyze energy data trends. Used Swift, API, and Shinobi for the graphics. Designed a system to forecast solar plant energy with neural networks in Matlab.Implemented REST API, JSON and web service in C# to transfer data.Machine learning: automated processes to teach the chip CogniMem neurons to detect objects and recognize patterns.

Analyzed data for non-invasive heart monitor using Dactron software, Matlab, FFT, and Visual C# to capture data sampling rates Decomposed and reconstructed wavelet packet using dB6 (level 5) filter for the separation of S1 and S2 Performed industry comparisons using published research on spectral analysis of the heart sound Assisted in prior-art research in cardiovascular technology Presented research OnCardio research on cardiac image analysis, Kalman filter, smoothing and wavelet at the IEEE Engineering in Medicine and Biology Society 2008 Annual International Conference

Developed and implemented functions for an instrument Centar/Frontier which gives the user the capability to grind chips to a nanometer level for chip analysis failure. Implemented warning displays that identified when the door of the instrument was open by reading pin number Provided solutions for customers/production teams attempting to read old data from the database. Provided solutions for the application engineer regarding specific settings on how to set parameters for failure analysis recipe.

Developed and implemented scientific software for a medical device which includes image processing, FFT, using Labview 5/ 8 C and C++ Improved the overall data in order to have a better prediction of the result. Implemented which are used to evaluate unknown data. Evaluated different processors and boards (from TI and Analog Devices) in order to port our prototype to a handheld device.

Responsible for high-end 2-dimensional gel electrophoresis analytical functionality. Scientific universities (Lawrence Livermore) use the software to facilitate progress in their research Designed, documented, and implemented 2D analysis such as spot detection, boundaries, element comparison, histogram, and Dendrogram. Designed and implemented image processing such as gray related preprocessing (brightness, gamma, contrast…) filtering methods (Gaussian average, median,…) and OCR using libraries from a third party LeadTool and 3D viewing and rotation with Proessential. Implemented dynamic link library and inter-process communication between applications.

Optimized the code in Linux of a program running on the TMS32VC5420 Digital Signal Processor which allows a DSP to emulate Timeslot Memory hardware. The program stores serial data into buffers allowing the host processor (Intel StrongARM) to inject data into the serial data stream. Optimized and increased the numbers of ports on a card by 35%. Designed, and implemented a tool using C that captures data from the DSPs on a Natural Microsystems CG6000 Compact PCI card; DSP Capture; provides a tool for diagnosing voice quality problems. Real-time environment code composer, VXWork, and Linux. Implemented a tool in a real-time environment using C on Linux that displays signaling bits and alarm statistics from T1 trunks for maintenance purposes.

Developed and maintained real time Molecular Device's proprietary software package (SoftMax Pro) which controls, acquires and analyzes scientific data from Molecular Devices Absorption and Fluorescence spectrophotometer microplate readers. This system provides high performance bioanalytical measurement, designed to accelerate drug discovery and clinical research. The software is multi- platform PC/ Mac, with Metrowerk and VisualC++. Implemented I/O interface (RS232). Responsible for the data analysis modification plan and implementation, and validation test plan. Conducted data analysis meetings. Proposed to Marketing, the VP of Research, and the project leader, the different algorithm alternatives as well as real data analysis examples. Modified a parser which calculates the linear regression of the rate of a kinetic reaction. Implemented curve fitting, normalization, smoothing algorithms, and modified all embedded mathematical functions to accommodate time tagged data. Developed software for Flexstation fast Kinetic essay for high throughput drugs discovery.

Programmed with Visual C++ /C, MFC, OLE, GUI application development with Win16, Win32 API, SDK. Application for Microsoft Windows. Designed, implemented, coded and tested a full cycle of commercial software using MRS (Marketing Requirement Specification) and SRS (Software Requirement Specifications). Implemented superposition of different graphs, modified calculation of rate to accommodate rounding issues.