Hengyi Tang

Hengyi Tang Email and Phone Number

Deep Learning and AI Engineer @ Skyline Dev Labs
Los Angeles, CA, US
Hengyi Tang's Location
Los Angeles, California, United States, United States
About Hengyi Tang

About me:- Graduate in Electronic Engineering (Machine Learning and Data Science) at USC. Expected to graduate in December 2024.- Research assistant at USC Information Sciences Institute and USC Cyber-Physical System Design (DesCyPhy) Lab.- Masters Students Honors Program in Ming Hsieh Department of ECE.My superpowers: - 3 years skilled developer of machine learning algorithms and python programming.- Interested in natural language processing and computer vision, and hope to take any direction as part of my future career plan.- Effective teamwork and joint software development capabilities.- Familiarity with a variety of machine learning and deep learning models (CNN, RNN, GNN, LLama, Transforms, ViT, GAN, LSTM, XGBoost)My specialties and hobbies:- Painting and Chinese calligraphy- Photoshop and Graphic Design- Outdoor hiking and photography

Hengyi Tang's Current Company Details
Skyline Dev Labs

Skyline Dev Labs

View
Deep Learning and AI Engineer
Los Angeles, CA, US
Website:
isi.edu
Employees:
309
Hengyi Tang Work Experience Details
  • Skyline Dev Labs
    Deep Learning And Ai Engineer
    Skyline Dev Labs
    Los Angeles, Ca, Us
  • Usc Cyber-Physical System Design Lab
    Research Assistant
    Usc Cyber-Physical System Design Lab Jun 2024 - Present
    Los Angeles, California, United States
    - Leveraged the pre-trained Deberta model to generate node pseudo-labels for feature augmentation in Text-Attributed Graphs (TAGs), enabling GNN models (GCN, RevGAT, GAMLP) to effectively utilize LLM outputs and improve robustness against structural perturbations by over 22%.- Developed LFQM, a layer-wise quantization model purification method aligned with full-precision ones, reducing the risk of attacks in LLMs under quantization conditions backdoors (QCBs).
  • Usc Information Sciences Institute
    Research Assistant
    Usc Information Sciences Institute Aug 2023 - Present
    Los Angeles, California, United States
    - Developed an adaptive, robust edge system for deploying diverse NLP and CV models on distributed edge devices, ensuring model accuracy while reducing communication and inference costs by 23%.- Implemented PipeEdge, a distributed framework using pipeline parallelism strategy in Pytorch that fine-tunes traditional CNN (Resnet, VGG, Alexnet) and transformer-based models (ViT, BERT, DEIT, LLaMA) to enable arbitrary layer sharding for efficient distributed processing across edge devices.- Employed AdaptivFloat, a dynamic precision quantization technique that autonomously clips the dynamic range of neural network parameters for faithful quantization encoding under limited computing resources.
  • Jiangsu Advanced Memory Semiconductor Co. Ltd.
    Machine Learning Engineer
    Jiangsu Advanced Memory Semiconductor Co. Ltd. Aug 2021 - Feb 2022
    China
    - Developed a rapid preliminary screening system for wafer lithography yield using intelligent data processing via Ali Cloud, enabling efficient reading of raw data and writing of supplementary information.- Designed and integrated a hybrid model combining Resnet-50/101 and LSGAN to analyze electron microscopy photos, accurately detecting wafer lithography depth and regularity, and marking typical problem areas to improve manual screening efficiency.- Implemented performance optimization strategies using tools like Java Profiler and SQL Optimizer, significantly enhancing the system’s ability to manage datasets exceeding 1 million records.
  • Multi-Granularity Prediction To Short-Term Traffic Flow Based On Ml
    Research Assistant
    Multi-Granularity Prediction To Short-Term Traffic Flow Based On Ml Jan 2021 - 2021
    Xi'An, Shaanxi, China
    - Developed an LSTM model to predict San Francisco urban road traffic flow by 3 million data over 5 years and infer road damage in the next few years considering factors such as season, temperature, construction standards, maintenance cycles, etc.- Constructed a hybrid model combining SRCNN, SRGAN and ViT for super-resolution on surveillance images, followed by extraction and recognition of standard texts and numerical characteristics of license plates.- Created a set of pre-processing algorithms including improving XGBoost for data types pre-classification and expansion of data feature dimensions, designing a Butterworth low-pass filter to reduce noise and enhancing the classical K-Means clustering algorithm to better recognize outliers.

Hengyi Tang Education Details

Frequently Asked Questions about Hengyi Tang

What company does Hengyi Tang work for?

Hengyi Tang works for Skyline Dev Labs

What is Hengyi Tang's role at the current company?

Hengyi Tang's current role is Deep Learning and AI Engineer.

What schools did Hengyi Tang attend?

Hengyi Tang attended Usc Viterbi School Of Engineering, 长安大学.

Who are Hengyi Tang's colleagues?

Hengyi Tang's colleagues are Rayyan Ibrahim, Amy Feng, Marco Kleimans, Osaze Shears, Parthav Joshi, Zhiwei Liu, Travis Haroldsen.

Not the Hengyi Tang you were looking for?

Free Chrome Extension

Find emails, phones & company data instantly

Find verified emails from LinkedIn profiles
Get direct phone numbers & mobile contacts
Access company data & employee information
Works directly on LinkedIn - no copy/paste needed
Get Chrome Extension - Free

Aero Online

Your AI prospecting assistant

Download 750 million emails and 100 million phone numbers

Access emails and phone numbers of over 750 million business users. Instantly download verified profiles using 20+ filters, including location, job title, company, function, and industry.