Erica Ji Yuen

Erica Ji Yuen Email and Phone Number

Generative AI Research Scientist @ Databricks Mosaic Research
Cambridge, MA, US
Erica Ji Yuen's Location
Cambridge, Massachusetts, United States, United States
About Erica Ji Yuen

Research scientist at Databricks, focusing on RAG, LLMs, embedding and retrieval systems. I joined Databricks via the acquisition of MosaicML. I use my expertise in ML research, design, and business to build and drive industry adoption of the latest cutting edge AI technology. Versatile, resourceful, and empathetic - I have a demonstrated history of building unconventional bridges across different industries and teams - driving cutting-edge innovation and explosive startup growth. In 2022-2023, I spearheaded MosaicML's brand strategy as an authoritative and community-driven thought leader in the ML research ecosystem and across enterprise industry verticals to drive the business - going from our first customer to 20M ARR + a 1.3B acquisition by Databricks in 1 year - one of the fastest tech unicorn exits in history. Since acquisition, I have been contributing to our research efforts for RAG (retrieval augmented generation) and synthetic data to enable enterprise customers to adopt GenAI solutions for their custom use cases, with a focus on synthetic high quality realistic benchmarks. Before that, I designed, launched, and scaled products in the AI Education space.Prior to my life with AI startups, I studied computer science at MIT, with a focus on multimodal AI and human computer-interaction, the effects of AI interfaces on cognitive psychology, and how we can design effective products that utilize multimodal human data input (text, audio, gestures, sketch, vision, etc) via AI and signal processing. I TA'd graduate level courses under the EECS department - Intelligent Multimodal Interfaces and Computational Photography. I also spent some time as a freelance makeup artist and fashion photographer, where I taught myself the fundamentals of visual branding, professional media production, business development, and working with clients. My areas of expertise include: Generative AI research, LLMs, computer vision, synthetic data, human-computer interaction, creative technology, marketing, visual design, operations, and AI startups.

Erica Ji Yuen's Current Company Details
Databricks Mosaic Research

Databricks Mosaic Research

View
Generative AI Research Scientist
Cambridge, MA, US
Erica Ji Yuen Work Experience Details
  • Databricks Mosaic Research
    Generative Ai Research Scientist
    Databricks Mosaic Research
    Cambridge, Ma, Us
    View
  • Databricks Mosaic Research
    Ai Research Scientist
    Databricks Mosaic Research Jul 2023 - Present
    San Francisco, California, Us
    RAG & Compound AI Systems for Industry on the Mosaic Research team (acquired by Databricks) Retrieval, Embedding models, Evaluation, Deep LearningSurveyed enterprise agent workflow use cases across verticals (finances, manufacturing, retail, healthcare, life sciences, telecommunications, retail, ecommerce, etc) to scope realistic and hard benchmarks Design and build high quality generalizable synthetic data pipelines for training / benchmarking to simulate end-to-end enterprise agent workflow use cases across verticals (finance, manufacturing, etc). These benchmarks are utilized across Databricks product engineering teams, including Inference and VectorSearch to deliver the best data intelligence on customer data.Developed training recipes and synthetic training data pipelines for finetuning embedding models for domain specific RAG use cases, improving cost, latency, and performance of retrieval systems.https://www.databricks.com/blog/announcing-hybrid-search-general-availability-mosaic-ai-vector-search
    View
  • Databricks Mosaic Research
    Mosaicml (Acquired By Databricks For 1.3B) - Ml Research Communications
    Databricks Mosaic Research Jun 2022 - Jul 2023
    San Francisco, California, Us
    As an early-stage hire on the growth, I spearheaded creative strategy and operations for the exponential growth of the MosaicML technical community among ML researchers. This role was the ultimate combination of my creative and deep technical background, and my ability to explain the most technical concepts in simple ways.In a scrappy and resource-limited environment, I drove the explosive growth (10X) of our product revenue, open source adoption, and social media following. Over 1 year, we went from our first customer to a 1.3B acquisition by Databricks with 20M+ ARR (one of the fastest unicorn exits in tech history). I was responsible for explaining MosaicML's highly technical research in simple, creative, and intuitive ways - branding MosaicML as the authoritative thought leader in the LLM research space for both the AI research community and the C-suite audience. Topics included distributed training, training LLMs from scratch, ML efficiency, diffusion models, and embedding models. My day to day looked like working closely with the CEO, CMO, CTO, and Chief Scientist to oversee the launch of our major research results - including telling stories about our research and engineering on our blogs, creating original technical illustrations, visual branding, social media growth strategy, strategizing event presence at ML academic conferences, and developing research partnerships to push the boundaries of large-scale machine learning. My illustrations have been featured in the press (Forbes), become viral on social media with millions impressions.Launches include: - MosaicML Platform (20M+ ARR) - MPT-7B, MPT-30B open source LLMs (3M+ downloads) - Training LLMs with AMD GPUs- MosaicML Inference - Training Stable Diffusion <50k with MosaicML- StreamingDatasets- MosaicBERT- MosaicResnet- Mosaic Segmentation models
    View
  • Inspiritai
    Technical Product Manager - Ai Education And Data Science
    Inspiritai Dec 2020 - Mar 2022
    - Designed technical AI & machine learning curriculum from scratch for 10+ courses customized for K12 and higher education covering topics such as computer vision, natural language processing, and data science. - Designed interactive lectures and developed Google Colab notebooks walking through students on preprocessing, training, and evaluating models across various real world applications with AI + Creativity, AI + Healthcare and more. - Courses launched had about 1M+ ARR within the first year. - Partner with stakeholders (schools, investors, businesses, families, students, instructors) to shape new initiatives based on daily feedback- Create, track, and present key progress indicators, project roadmaps, milestones, risks, and outcomes to hit performance targets- Speak at public hackathons, schools, and education conferences- Instructed 200+ 2.5 hour AI sessions, adapting to the needs of an audience ages 8-25, class sizes of 2-100- Oversaw the hiring and training of 500+ instructors and scaled processes to manage instructional quality - Ideated growth and launch strategy for new endeavors aimed at rapidly scaling in the EdTech space
  • Mit Media Lab
    Ai + Human Computer Interaction Researcher
    Mit Media Lab Mar 2019 - May 2020
    Cambridge, Massachusetts, Us
    I investigated how multimodal generative AI can be a tool to encourage creative psychology and divergent thinking. I was particularly interested in how new AI driven interfaces can be assistive tools for those experiencing mood disorders, particularly depression and ptsd. I worked with vectorized, rasterized, time series, and language data in this project. This interface took human sketch data, classified and colorized it in real time, and encouraged creative story telling by recommending new elements to add using embedding models. I collected training data for sketch to cartoons, trained pix2pix GAN model. I developed a web app (fullstack) Flask and Javascript to perform live sketch recognition on digital strokes, modeling the context of user drawings to adaptively colorize sketches using the GAN models in real time. In addition, I developed features to allow users to drag and drop, transform, and rotate elements on their canvas. Designed and conducted qualitative human study with participants to analyze changes in creativity from tool, using mixed methods. I recruited 25+ participants, and users reported that features like NLP-based associative word generation + live adaptive colorization provided additional assistance with divergent thinking compared to other illustration tools.
    View
  • Massachusetts Institute Of Technology
    Graduate Teaching Assistant
    Massachusetts Institute Of Technology Jan 2019 - May 2020
    Cambridge, Ma, Us
    Teaching assistant for graduate level courses Computational Photography and Intelligent Multimodal User Interfaces.Prepare and administer coursework, grade psets, and mentor 300+ students to learn the material.Topics covered in Computational Photography (taught in C++):Denoising and Demosaicing, Convolution and Bilateral Filtering, Resampling, Warping and Morphing, Harris Corners, Features and Automatic Panoramas, Image Processing Optimization using HalideTopics covered in Intelligent Multimodal User Interfaces (taught in Python and Javascript):Sketch Understanding, Body Tracking and Gesture Recognition, Face Recognition, Speech Understanding, Affective Computing, Cross Modal Understanding
    View
  • Samsung Electronics
    Machine Learning Engineer + Product Design
    Samsung Electronics Jun 2019 - Aug 2019
    Suwon-Si, Gyeonggi-Do, Kr
    During my internship at Samsung Design and Innovation Center (SDIC), my project was perform research on how Samsung can develop the next generation AI Camera for their flagship phones. SDIC was primarily a design studio with UX and Industrial Design teams, and they were excited to see collaborate with someone with a machine learning background to innovate their products. I reported to the DIrector of Engineering at SDIC, and I presented my work to the Director of SDIC and Executive VP of Samsung - and two years after my internship the features I designed were released in the Galaxy S21.I worked with saliency and segmentation models, object detection models, optical flow analysis, multimodal signals processing for video, image, and audio, and computational photography methods. I independently designed the UX and prototyped first working versions of Portrait Mode Life Focus "Color Point" feature in Galaxy S21, which detects the subjects using AI and makes the background black and white. I created this idea and worked the first working implementations the end to end experience, integrating AI models for detecting saliency and image segmentation from research papers. Drawing from my experience as a professional fashion photographer, my original idea was inspired by helping creators and small businesses create better product photos using AI. I tested my original prototype across different scenarios with clean results. - "Single Take" feature on the Galaxy S21, which generates several "best shots" from a 10 second video. I created prototypes that generated "best shots" based on changes in sound and optical flow, with integrated auto-cropping, video stabilization, and object trackingI contributed towards developing user journeys for audio experiences for the Galaxy Buds, as well as in-store customer experiences. I developed roadmap for potential camera features, analyzing tradeoffs between UX and mobile computational performance
    View
  • Ey Studio
    Freelance Fashion + Beauty Photographer And Makeup Artist
    Ey Studio Jan 2018 - Jun 2019
    Freelance beauty and fashion photographer and makeup artist. I also specialize in hyperrealistic gore and special effects makeup. These experiences were my first introduction to building a business, professional design, and branding. I also developed an understanding of industry-best practices of using software and tools for media and fabrication. I am passionate about capturing breath-taking imagery and telling stories. As a makeup artist turned photographer, I pay extra attention to detail and creativity when styling and shooting. I work both independently and collaboratively when curating my creative vision on set. My process consists of understanding the mood of the imagery requested by the client, and bringing their vision to life through styling, light modification, and retouching in a timely fashion.Past clients include independent fashion designers and Jasz Couture, an evening gown brand specializing in pageant and prom gowns.
  • Microsoft
    Software Engineer Intern
    Microsoft Jun 2018 - Aug 2018
    Redmond, Washington, Us
    Collaborated closely with UX designers to improve the Intune Company Portal iOS app experience of browsing and installing enterprise resources, raising App Store user rating from 2 to 4.5 stars after shipping new featuresImplemented and shipped browsing features using XCode and SwiftFeatures built currently displayed as main product screenshots on the iOS App Store
    View
  • Appinventor - Mit Csail
    Undergraduate Researcher - Superurop
    Appinventor - Mit Csail Sep 2017 - Jun 2018
    Created data visualization features for mobile app creating platform using Java, d3.js, HTML, and CSS.AppInventor aims to provide accessible and easy-to-learn mobile development education for beginner programmers.
  • Microsoft
    Software Engineer Intern
    Microsoft Jan 2018 - Jan 2018
    Redmond, Washington, Us
    Created specs, and implemented grid, snapping, and layout constraint features with C# for Ink to Code, a platform for users to quickly prototype Windows and Android apps with runnable code from gesture recognition with the Microsoft Surface Pen - worked with a team of 5
    View
  • Intuit
    Software Engineer Intern
    Intuit May 2017 - Aug 2017
    Mountain View, California, Us
    Rapidly prototyped new features using Vue.js for UX research to identify growth potential for new product,Turbo, tracking analytics to understand what personalized insights users want to better manage their financial futureDeveloped API to load dynamic personalized user content to Turbo based on various data sources using Java Spring MVC
    View
  • Mit Media Lab
    Hci Researcher Assistant For Duoskin
    Mit Media Lab Jan 2016 - May 2017
    Cambridge, Massachusetts, Us
    Worked under the Living Mobile group to design and fabricate functional conductive temporary tattoos with embedded circuits and improved application methods using experiences with cosmetics.My design, a light up tattoo necklace, ended up being the main image used when the project got featured on Forbes, CNN, the front page of reddit, TechCrunch, etc.These DIY flash tattoos, made of gold leaf, have the ability to conduct electricity. I played an active role in combining the aesthetics of the tattoos with my electrical engineering experience to create new applications and interfaces, including skin touch sensors and light up tattoo necklaces. My makeup artist experience helped make the fabrication process more robust and efficient.I designed applications to enable users to control mobile devices viatouch gestures on their skin, Bluetooth, and near field communication. Used Arduino and Processing to create touch interaction visualizations.I collaborated with MIT and Singapore University of Technology and Design to teach design workshops using Duoskin.2017 SXSW Interaction Innovation Award; exhibited at New York FashionWeek; widely featured in the press (CNN, Forbes, Ars Technica)
    View
  • National Geographic Learning
    Virtual Reality Development Intern
    National Geographic Learning Jun 2016 - Aug 2016
    Utilized Maya, Unreal Engine, and Oculus Rift to create a virtual reality simulator for the Canadarm2, a robotic arm on the International Space Station3D modelled assets using Maya and Photoshop for a VR animated film about climate change, the Boy and the Cloud
  • Educatrium
    Front End Web Developer
    Educatrium Jun 2015 - Aug 2015
    Used HTML5, CSS3, and jQuery to design and launch company’s website

Erica Ji Yuen Skills

Powerpoint Algorithms Teaching Web Design R (Programming Language Adobe Photoshop Cascading Style Sheets Html Artificial Intelligence Programming Java Microsoft Word Cosmetics Javascript C++ Python Research Computer Science Software Development Front End Development C# Html5 Maya

Erica Ji Yuen Education Details

  • Massachusetts Institute Of Technology
    Massachusetts Institute Of Technology
    Computer Science
  • Massachusetts Institute Of Technology
    Massachusetts Institute Of Technology
    Computer Engineering
  • Winchester High School
    Winchester High School
    High School Diploma

Frequently Asked Questions about Erica Ji Yuen

What company does Erica Ji Yuen work for?

Erica Ji Yuen works for Databricks Mosaic Research

What is Erica Ji Yuen's role at the current company?

Erica Ji Yuen's current role is Generative AI Research Scientist.

What schools did Erica Ji Yuen attend?

Erica Ji Yuen attended Massachusetts Institute Of Technology, Massachusetts Institute Of Technology, Winchester High School.

What skills is Erica Ji Yuen known for?

Erica Ji Yuen has skills like Powerpoint, Algorithms, Teaching, Web Design, R (Programming Language, Adobe Photoshop, Cascading Style Sheets, Html, Artificial Intelligence, Programming, Java, Microsoft Word.

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.