I am the CTO of Personal AI, a rapidly growing start-up building PLM (Personal Language Model) or "Digital Twins" for subject matter experts, lawyers, healthcare practitioners, financial analysts, executives, educators, and thought-leaders. Our customers are enterprises and organizations who want to leverage their executives' deep subject matter expertise by making their Digital Twin available 24x7x365, and easily accessible via chat, email, SMS, voice, and other communication mediums.

The "Lovelace Collective" is a platform that amplifies the stories of diverse and under-represented founders and operators. Inspired by the remarkable Ada Lovelace, we celebrate individuals who are brilliant and committed to their passions, yet have often felt hidden and alone in their journey. Join us to explore the power of the collective support and strength that comes from knowing you are never alone.🎧 https://www.thelovelace.co/listen

Follow me @ https://marblecollective.com/sharon.zhang

• Analytics for profiling and optimizing low latency system with multicore/multithreaded architecture solutions using C++• Research and implementation of trading algorithms to improve scalability and profitability of high frequency trading (HFT) system• Data analytics and metrics derived from large amounts of production training data• Infrastructure, production and operation automation for improving robustness of production system; testing and releasing on tight production cycles

ASR Applied Research Group: • Developed and compared multiple clustering techniques to retrieve additional data for creating hybrid models from both speaker dependent and speaker independent data• Improved document structure detection and text formatting models for creating structured outputs from unstructured recognition input• Created infrastructure for research farm queuing and utilization; constructed standard train/test sets; developed golden corpus and evaluation analytics• Formulated and executed detailed experimental and test plans• Released and monitored rollout for model building packagesClinical Language Understanding Group (Founding NLP):• Compared heuristics, VSM and language modeling techniques to create section specific modeling for topic detection for EMR• Created dictionary from the SNOMED ontology for use in machine learning models and built language models to do dictionary cleanup and pruning• Employed Active learning techniques to select high relevant data for annotation for mention detectionQuestion Answering Group (Watson MD):• Extracted information from semi-structured sources (i.e. CDC website, medical texts, ontology) and converted them into structured RDF formats for information retrieval• Built Indri indices from medical sources for free text answer retrieval• Ran question answering experiments and evaluated retrieval accuracy on individual sources• Retrieval error analysis and evaluation, implementation of features as solutions

Research and Development in HealthCare division, focusing on using machine learning and statistical natural language processing methods to generate Medical Records

Jan. 2005 - Jun. 2006 - Sloan School of Management- Market data mining, analysis and research for Biotech firms specialized in cancer drugs.Jun. 2006 - Aug. 2006 - Women's Technology Program- TA and Residential Tutor for Discrete MathematicsAug. 2006 - Dec. 2006 - BioMEMs Group / RLE- Experimented with microfluidics and stem cellsApr. 2007 - Jun. 2008 - Clinical Decision Group / CSAIL - Worked on information extraction and information retrieval from EMR using ML/NLP methodsSep. 2007 - June 2008 - Consultant @ IS&T- Worked at HelpDesk to support MIT affiliates. Main responsibility includes diagnosing and resolving system, application, network and other software related problems.