Claire Li Email & Phone Number

Technical Co-Founder, and CTO of AI Bio Check, a seed-stage machine vision robotics startup recently empowered by a significant $500k funding from the HF0 fund.I serve as the Chief Technology Officer. (ﾉ´ヮ`)ﾉ*: ･ﾟOur mission is to redefine lab automation through advanced AI and robotics, with the goal of enhancing efficiency and accuracy in biotech research and development.* Architecting and overseeing the development of AI Bio Check’s machine vision robotics platform, leveraging proprietary AI/ML technology to automate and optimize laboratory tasks previously bound by manual limitations.* Strategic Leadership: Charting the technological direction of AI Bio Check, aligning product development with strategic business goals, hiring engineering talent, and fostering a culture of innovation and excellence within the first engineering team* Security and Scalability: Implementing a proactive security posture, drawing on my extensive background in cloud security and DevSecOps methodologies. My approach ensures that every component of our robotics systems is built with the highest security standards to protect sensitive data and maintain trust in our automation solutions* Overseeing the creation and implementation of robust operational processes, from microservice architecture development to streamlined CI/CD pipelines, ensuring efficient deployment and operational scalability across our infrastructure* Product engineer #0, architected the MVP before raising capital, developed a FULL-FULL stack platform that spans from the embedded software of the robot's hardware, to robust AI/ML infra architecture to support real-time inference, cloud training pipeline, front-end and back-end of SaaS web application for training, labeling data, and controlling the robot. A challenge that has amalgamated many facets of my ability as an engineer, technical leader, and culminated in our flagship robot and platform!

Santa Clara, California, Us

As the resident cloud security infra/automation expert, I led multifaceted initiatives across the infra stack and software deployment lifecycle, following a devSecOps methodology. My contributons included:*Designed and implemented security controls and scalable IAM privilege designs, strengthening platform security while advocating secure tooling. Proficient in automating security operations and fostering collaboration with engineering teams to establish a golden standard.* Developed microservice for consolidating malware family names between our platform and antivirus team * Enhanced deployment efficiency and cloud infrastructure security with upgraded or newly created IaC components, including Terraform modules and Ansible roles. Implemented encryption for GKE clusters and GCE controller VMs. Introduced fine-grained IAM design to control deployment mechanisms based on engineers' LDAP group membership.*Led optimization efforts for PANW's malware detection machine learning services within GKE, emphasizing automated horizontal scalability and performance.*Developed novel in-house Python utility for GCP Monitoring Dashboard preservation, exporting resources, and generating valid Terraform code for deployment*Designed and implemented a streamlined CI/CD pipeline for PANW's malware detection microservices, optimizing deployment across various environments, including government FEDRAMP secured setups. Authored CI/CD logic, resulting in a 38% reduction in execution time.*Upgraded a legacy system for deploying an anti-virus data API, transitioning from shared server wildcard certificates to environment-specific certificates, enhancing security and certificate management.*Automated and standardized SSL load balancer certificate renewal across multiple environments, reducing errors and improving certificate management for secure API communications. This architectural automation significantly enhanced efficiency, particularly in production environments.

Santa Clara, California, Us

Advanced malware analysis engine, threat detection, deployment & operation of infrastructure for a cloud-based firewall on a massive scale!!!! ✧・ﾟ:*(´ ꒳ ` ✿)*:・ﾟ◦ Wildfire SRE team. Served as SRE lead twice. For bringing Wildfire up into UK, Australia, to support sample analysis in the quantity of 300,000 files and upwards per day for our clients, biggest client was the British government's National Health Service. Our engine handled computer traffic in public hospitals, and other critical public services, during the global pandemic.◦ I am the team’s domain expert in cloud security, I designed custom solutions for many of Info Sec’s requirements, and have implemented them across US, UK, Canada, Australia. ◦ Designed Terraform modules for onboarding engineers into GCP WF Cloudecosystem, controlling IAM privileges. Became team expert on GCE OS Login, and wrote custom programs for implementing OS Login, IAM driven-SSH-ACL to VMs. ◦ Designed Ansible & Python module solutions for service account and RSA key pair binding, introduced group-service account architecture to the team, group accounts that have curated privileges for deployments.◦ Strong skillset in automation, and built two generations (Jenkins, and Gitlab CI) of CICD pipelines for highly available threat intelligence service.◦ Wrote extensive security SOP (standard operating procedure) for the team, and enforced secure practices compliant with Info Sec. ◦ Reduced engineering man-hours from 3 weeks to 18 minutes, for a fully automated, end-to-end threat intelligence build and GKE deployment of 15 microservices. Designed the architecture for cross-global WF environments. Automated the Helm chart preparation, docker builds, IAM policy for auth/pushing/syncing to container registries, and final Helm deployment to GKE cluster for any amount of GCP regions.

Cupertino, California, Us

◦ Anti-Fraud engineering team, SRE (service reliability engineering), worked on horizontal and vertical security infrastructure problems for Apple (including iCloud, iTunes, iMsg). Responsible for capacity planning, scaling, deploying & securing data scientist’s Spark & Scala machine learning models for fraud detection.◦ Maintained and built new features for Spark system that will deploy a Spark job (by executing an Apache Spark-Submit in the context of SupervisorD) onto a designated cluster. Brought up bare-metal QA host cluster for system, in addition to a “gateway” host which manages SupervisorD, downloads config files & jars from external environment, deploys and confirms success/failure.◦ Wrote Ansible playbook features for manual “gateway” selection for user through .yml configs, or passed into the command at runtime. Wrote features for overriding previous Spark deployments, programmatic supervisor config cleanup.◦ Created a production freeze mode, where we can block automated Spark deployments to production during critical company announcements, while allowing QA and validation deployments to continue without intervention◦ Set up telemetry monitoring & health checks for Spark system, with automated hourly tests that detect any breakpoint in the pipeline, and gateway host errors◦ Worked with the team on mass Spark job (and dependencies) migration between environments while ensuring minimal impact.◦ Worked on SRE operations such as 24/7 on-call, upgrading environments for large fleet of hosts, monitoring host/app health, triaging/resolving errors on the application & host level, identifying & disabling insecure applications/ports, fixing SRE tools such as automated Ansible role updates, and onboarding secrets (mutual auth keys, passwords) into Vault◦ Drove initiative for defining secrets standards (hierarchy), and designing an internal REST API for SRE team to list policies, modify secret & policy relationships, store secret assets programmatically

Sunnyvale, Ca, Us

◦ Monitoring Infrastructure team, worked on Squiggles, a high-performance data rendering library for time-series. Library would display server and system metrics across LinkedIn, inside of monitoring application, inGraphs. ◦ Responsible for micro-optimizations in library dataset sorts to reduce time and space complexity, re-architecting Squiggles to support additional graph types that can be rendered at a highnumber of query requests and traffic throughput◦ Developed multi-threaded solution for library to offload heavier graph computations to worker threads, was able to demonstrate changes from ~6900 milliseconds to 700 milliseconds, resulting in a ~90% increase in run-time efficiency◦ Converted existing time-series graph types into new graph type, "horizon graphs", to be available for inGraphs users across LinkedIn. Graphs allow higher data density along more granular time invervals, which still preserving graph resolution. Worked on preparing the newest version of the library for open source publication

San Diego, California, Us

◦ Developed for the Stallion project at Epsilon Systems, for client SPAWAR (Space and Naval Warfare Systems Command)◦ Systems data mining application for parsing communication logs between naval destroyer transponders, and calculating the TFIDF value (term frequency–inverse document frequency) for certain target terms. Ran across a Hadoop cluster, and interacts with Apache Spark. ◦ Responsible for interpreting resilient distributed datasets, and writing user-facing application in Scala◦ Worked on project for SPAWAR division in the Navy for ship simulations, and scaling existing application. Transitioned application to Apache Apex, ensured higher through-put and fault tolerance for big data application

Santa Clara, California, Us

**TECHNICAL DETAILS for both projects can be viewed below under "PROJECTS" on my profile **◦ Rehired for second consecutive application development internship with Intel Corporation, at headquarters.◦ Development of enterprise software for Intel Corporation, for internal use. Project 1 Overview:Previously: Outside customers had no access to updated shipping information for samples in transit and would have to go through a point-of-contact in Intel, then a member of a CRM team that would have access to shipping information. ◦ I developed a feature that would automate this process and give shipping status information to the customer in the UI of our internal application.◦ Queried a data virtualization platform via web service to check for hardware samples in Intel headquarters in transit◦ Formatted XML data in an advanced SQL query that would store unique data per call, in a Temporary Table that would only exist for that session. Data would then be passed to a UI (modal popup) in JavaScript.◦ This cut down massively on the amount of memory consumed, because we wouldn’t have to continuously update SQL server and write information to the database multiple times for a shipment which has a short lifespan. This also saved a lot of man-hours within Intel, because our internal application is used by over 1000 employees internallyProject 2 Overview:Created a RESTful API from scratch in ASP.net, and a UI that would consume my web service to be used by project managers who would have access to a set of different data sheets, and overwrite each other’s changes. I devised a way to search, read and write these data files with certain restrictions.◦ My inspiration behind this RESTful API was to create a read-write utility with similar behavior to a Relational Database, given to end-users with no experience in SQL or database administration. My goal was to treating CSV documents as a restricted “back-end” that could only be modified via my service.

Santa Clara, California, Us

◦ Development of enterprise software for Intel Corporation, for internal use◦ End-to-end production, testing, in Agile & SCRUM workflow◦ Developed a system alerts feature for an internal application that faces over 1,000 end-users inside Intel. Allows a system administrator to schedule maintenance & down time for the internal application. Information is sent to SQL Server, and queried by time-stamp to determine when to publish message. ◦ When an alert message is due, the error message is published in a UI element using jQuery◦ Worked on Selenium/ FitNesse framework for automation testing of the utility◦ Wrote a fuzzy search algorithm for sorting large data sets, and determining the string metric difference between user input, and a large data set of over 400 entries of Intel customer information, and instantly producing search query suggestions or auto-complete◦ Predicts when user is done typing, or makes appropriate adjustments through a series of AJAX calls, and a JavaScript delay timer◦ Formats large data sets into JSON array from a local directory, and strips entries of special characters and spaces◦ Determines which strings out of the large data set match what the user is typing, and sends the top matches to a drop down menu below the search input, using JavaScript◦ Average execution time for each AJAX call is consistently below 150 milliseconds, or 0.15 seconds, for a string matching against a test data set of 480 entries

Norwalk, Connecticut, Us

◦ Interned for city government, and Xerox. Wrote a "Riverside City Open Data Permit Loader" application◦ WIN32 application in Python that queries an Oracle database for all building permit information within a certain time range for the city of Riverside. Using the cx_Oracle driver, the application retrieves the SQL data then converts to JSON format to programmatically append to an open web transparency portal for the city◦ Generates an HMAC, encrypts the data in md5, and commits the data through an HTTP Post Request◦ Program can be either run in the console through a daily CRONtab and Shell file, or manually run via GUI desktop application written using Python’s tKinter graphics library◦ Worked on demo mobile GIS application in Java for Android◦ Application utilized Google Maps API and would mathematically calculate a “historic crime coefficient” to determine a level or danger in the immediate physical location. The UI would notify you if you were standing a green, yellow, or red zone. All calculations were generated through server-side scripting in PHP, and would pass the data to the Android application◦ Application could sort crime data based off date, immediate location, other addresses, or time of day

San Diego, Ca, Us

◦ Developed MySQL database for company’s information, along with local UI submission form that ran off server- side PHP scripting, to use for database updates◦ Built file-directory crawler using a combination of batch scripting, transferring the file information into Microsoft Excel, and Oracle Workbench. Used for quality assurance during the mass file migration◦ Communicated with other IT workers contracted by Open Mobile Alliance in the UK◦ Configured PHP BB forum for the company and its partners to use, administrator. Front-end development to adjust the UI.

◦ Marketing rewards program start-up in Carlsbad, San Diego◦ Worked closely with marketing director, Bob Krause, on graphic design, material for sales pitches, front-end web development. Built several demo websites, using HTML, JavaScript, CSS. Responsible to migrating development files to production environment through FTP client