Help & Supportcareers@gravityer.com
AI is analyzing your overall score…
Identifying your key strengths…
Evaluating your skill match against the job requirements…
Assessing your cultural and operational fit
Senior Machine Learning Engineer at Apple • Apple Intelligence Foundation Models • Multimodal LLM Engineer - audio/speech
Working on Apple's multimodal foundation models powering Apple Intelligence. Combining research rigor with production-grade engineering, designed and executed the full program from data curation/CPT/SFT through reinforcement learning on billion-parameter speech-language models. Identified key accuracy gaps in multimodal speech models and proposed a novel pretraining approach to address it. Proficient in conducting research and production oriented ASR, NLP, LLM Deep-Learning modeling R&D workflows and big data (Spark) pre/post-processing pipelines. Solid in-depth theoretical backgrounds and source-code-debugging experiences of Transformer, GPT, BERT, RoBerta for ASR Language Model (LM) and Acoustic Model. Significant contribution to the organization's system and metric improvements like WER, Cross Entropy and F1 score. Specialty in Entity-centric ASR/NLP problems, and attention based models like Transformer, NNLM, chatGPT-like models. Previous project experiences on various topics like NER, entity expansion, entity correction and machine translation. I am a Senior Machine Learning Engineer with years of project leadership experiences in conducting deep learning/statistical modeling and Automatic Speech Recognition (ASR) on large-scale datasets and with both industrial and academic research experiences for cutting-edge applications of machine learning algorithms and deep learning frameworks. Over the past few years I have directly driven and led LLM projects for improving entity-centric speech utterance recognition accuracy, as well as implemented various Language Model (LM) hacking to improve WER, including Transformer based ASR rewriter, GPT2 word piece Language Model reranker, LSTM/CNN based NLP/NLU related models for various topics with internal tools as well as deploy large-scale data pipeline automation with Spark and Airflow for metrics and features . Befo
Duke University
Master’s Degree, Statistical Science
January 1, 2015 – January 1, 2017
UCLA
Mathematics and Statistics
January 1, 2012 – January 1, 2013
Fudan University
Bachelor's Degree, Economy
January 1, 2010 – January 1, 2015
Fudan University
Bachelor's Degree, Sociology
January 1, 2010 – January 1, 2015
Apple
Senior Machine Learning Engineer, Siri Speech
June 1, 2017 – Present
Cupertino, CA
Spectral MD (Algorithm Group)
Summer Intern -- Deep Learning and Computer Vision
June 1, 2016 – August 1, 2016
Dallas-Fort Worth Metroplex
Duke University
CAH Research Assistant
March 1, 2016 – May 1, 2016
Raleigh-Durham-Chapel Hill Area
Fudan University
Research Assistant -- Social Network Analysis
February 1, 2014 – November 1, 2014
Shanghai, China
Nielsen Company
Data Analyst (machine learning)
November 1, 2013 – June 1, 2014
Shanghai, China
Industrial and Commercial Bank of China
Quantitative Analyst
August 1, 2013 – October 1, 2013
Shanghai, China
UCLA
Research Assistant -- Ranking Algorithm
June 1, 2013 – August 1, 2013
Los Angeles
Tongji University
Research Assistant: Regression Analysis
July 1, 2012 – September 1, 2012
Shanghai, China
AIESEC
Team Leader
March 1, 2011 – June 1, 2012
Shanghai, China
Advanced machine learning – Topic modeling and clustering
January 1, 2017 – Present
• Natural language processing (NLP) by comparing Bayesian Nonparametric methods and Deep Neural Networks • Researched on generative models like LDA, Dirichlet Process, Chinese Restaurant Process, Indian Buffet Process, etc.
Kaggle machine learning competition – Regression prediction on severity
November 1, 2016 – December 1, 2016
• Constructed machine learning regression models to make accurate predictions on severity of insurance claims • Achieved top prediction ranking by model selection among regressions, XGBoost and neural networks using Python
Stock Market Prediction - An empirical comparison between Time Series Analysis, Hidden Markov Model (HMM) and Artificial Neural Network (ANN)
May 1, 2016 – Present
• Implemented Time series analysis on stock prices scrapped from Yahoo Finance • Compare Time series models with Hidden Markov Models and Neural Network • Proposed investment strategies based on the prediction results.
Twitter data visualization and Shiny web application developement
December 1, 2015 – Present
• Coordinated as the core designer of the Dashboard web application for interactive data visualization using R • Scrapped Tweets using Twitter API for text mining • Developed descriptive statistics of the Tweets • Visualized statistical analysis and data in Shiny web application in R
Reddit Big Data text mining project
November 1, 2015 – Present
• Designed and developed analytics in Hadoop MapReduce and detected top monthly trends by using Rhipe • Utilized Spark for information retrieval and conducted text mining
Distance analysis on Web-scrapped data
October 1, 2015 – Present
• Scrapped spatial data using website API or by inspecting website elements • Conducted distance analysis based on the spatial data
SAS Certified Base Programmer for SAS 9
SAS
June 24, 2026 – Present
SAS Certified Advanced Programmer for SAS 9
SAS
June 24, 2026 – Present
Cultural Fit Analysis
The candidate's diverse project portfolio, ranging from stock market prediction to social network analysis and big data text mining, indicates a broad interest in data-driven problems. Experience in both academic research and industry (Apple, Nielsen) suggests adaptability to different work environments. The variety of tools and techniques used (R, Python, Hadoop, Spark, SQL, various ML/DL frameworks) points to a willingness to learn and apply new technologies, which aligns well with a dynamic data-focused culture. The target role of 'Data Analyst' aligns with the candidate's core competencies in data analysis, modeling, and interpretation, although their recent experience at Apple as a 'Senior Machine Learning Engineer' suggests a more advanced, specialized ML focus than a typical Data Analyst role.
Soft Skills & Operational Fit
The candidate's project and work experience descriptions suggest strong analytical thinking, problem-solving, and a research-oriented approach. Roles like 'Team Leader' at AIESEC indicate leadership and organizational skills. The diverse project portfolio demonstrates adaptability and a proactive learning attitude. However, direct evidence of collaboration style or stress handling is not explicitly detailed in the provided data.