AI Engineer on the LexisNexis Agentic AI Team (more details coming soon)

Teaching Java, JavaScript, Python, and Machine Learning to students ages 12-16 via online private classes + summer camps.

Enhanced an internal agentic AI platform that automates up to 95% of the SDLC for enterprise ELERA workflows.

Spearheaded the design and implementation of a RAG-based document ingestion pipeline using LangGraph and Weaviate, improving knowledge retrieval accuracy by 30%.

Streamlined stages from requirements to deployment using Dockerized microservices, AWS, and Jenkins CI/CD pipelines.

Built a QA Agent for autonomous testing with 6 new task types (feature, release, service pack test plan & test cases generation), plus BRD analysis and gap checks, with deep integration into Jira, GitLab, JUnit, Selenium, Cypress, and TestRail.

Developed 4 multi-agent workflows using Kafka, Redis, LangChain, LangGraph, Neo4j, Weaviate, and RAG.

Shipped an admin stats dashboard with real-time WebSocket monitoring, system health metrics, and MongoDB aggregations in a responsive Vue UI, enhancing operational visibility.

Implemented a comprehensive AI cost analytics system with multi-provider pricing (OpenAI, Anthropic, Gemini), tiered models, per-task attribution, and real-time visualization, boosting cost transparency by 55%.

Delivered an API documentation platform with Swagger UI + ReDoc, auto-generated OpenAPI specs, and JSON/YAML export.

Resolved 15 critical navigation and UX defects, including an XSS fix, to improve platform stability and security.

Collaborated in an Agile environment, participating in daily stand-ups, sprint planning, and code reviews to ensure timely delivery of high-quality features.

🚀 Received return offer ✨

Tech Stack: TypeScript, Python, Express.js, Java, Spring, Docker, Kubernetes, AWS, Azure, Jenkins, Kafka, Redis, LangChain, LangGraph, CrewAI, Neo4j, Weaviate, Chroma, RAG, MCP, Vue.js, MongoDB.

Served as a Teaching Assistant for COMP-426: Modern, Full-Stack Web Development at The University of North Carolina at Chapel Hill, under the guidance of Professor Ajay Gandecha.

Assisted in teaching advanced web development concepts, including TypeScript, React, Node.js, Next.js, Express.js, PostgreSQL, and Supabase to a class of 200 undergraduate students.

Conducted weekly 8+ office hours to provide hands-on assistance and clarify complex topics.

Graded assignments and projects, ensuring timely and constructive feedback to help students improve their coding skills.

Collaborated with the lead instructor to develop course materials and improve the curriculum based on student feedback.

Helped students build full-stack web applications, fostering a practical understanding of modern web technologies and best practices.

Worked on the FRG Visualization of Risk (VOR) platform, reducing UI load times by 20% and increasing Django API throughput by 15% through targeted Angular refactoring and backend optimizations.

Integrated a RAG-based, context-aware chatbot using FAISS vector stores, LangChain, LLaMA via Transformers pipelines, NgRx, and Apache Kafka with Zookeeper for streaming real-time document updates to keep the AI’s knowledge up-to-date.

Implemented an RLHF feedback loop to continuously improve chatbot accuracy based on user interactions, increasing response accuracy by 20%.

Developed an AI Logs Explainer that transforms 500+ complex risk-engine system logs into clear, human-friendly summaries, cutting manual analysis time by 70%.

Built a Performance Analyzer that combines LLM insights with RAG-enriched database queries, raising query efficiency by 25% and delivering 100+ actionable recommendations/month.

Built a Forecasting Module powered by Facebook Prophet, achieving 92% accuracy on historical data.

Created an Event Predictor that turns free-form user descriptions into quantified scenario impacts on demand, serving 200+ scenario analyses/month.

Delivered a conversational Model Explainer to demystify economic model structures and behavior within token-limit constraints.

Designed interactive sensitivity and impact visualizations, including a coefficient-product interaction matrix, for deeper risk analysis.

Wrote Ansible playbooks and roles to automate infrastructure provisioning and deployment of services across multi-environment AWS clusters.

Contributed to comprehensive Playwright end-to-end test suites automating UI workflows across Angular components, including user authentication, data-driven visualization interactions, and error-state validations.

Integrated tests into the CI/CD pipeline with parallel execution, cross-browser coverage, and detailed reporting to catch regressions early.

Gained experience with WSL and Ubuntu.

Tech stack: AI/ML, Angular, TypeScript, NgRx, Syncfusion, Material UI, Django & Django ORM, Python, Go, PostgreSQL, MSSQL, AWS, Apache Kafka, Zookeeper, FAISS, LangChain, LLaMA, Prophet, Playwright, Ansible, HashiCorp (Vault, Consul, Terraform), Docker, WSL, Ubuntu, Shell scripting.

Designed and implemented a React / Next.js frontend delivering a responsive, accessible interface for digital health surveys.

Built and maintained a Python / Flask backend API to securely ingest, validate, and store participant data in PostgreSQL.

Deployed the entire stack to AWS (ECR → ECS → ALB) with RDS for persistence and S3 for static asset hosting.

Configured Celery + Redis for asynchronous task processing (e.g. reminder emails, bulk data exports).

Integrated JWT authentication and flask-cors for secure cross-origin communication with our Next.js frontend.

Automated infrastructure provisioning using Terraform, and container orchestration with Docker Compose in local development.

Wrote comprehensive unit and integration tests (pytest for backend, Cypress for frontend) to ensure reliability of every release.

Collaborated closely with health economists and policy researchers to translate study protocols into production-ready web features.

Tech stack: React, Next.js, Python, Flask, PostgreSQL, AWS (ECR, ECS, ALB, RDS, S3), Celery, Redis, Terraform, Docker Compose, JWT, flask-cors, pytest, Cypress.

Developed and maintained a React and Flask full-stack web application that allows the company's staff to manage and track their projects and clients.

Assisted in the development of a Flask & Python backend that processes user data and sends it to the database.

Designed an improved database architecture that increased the application's performance by 30% and implemented Role-Based Access Control (RBAC) with Auth0 to enhance security.

Enhanced application security by implementing OAuth 2.0 and JWT for user authentication, and measures to prevent CSRF and XSS attacks.

Technology Stack: React, Flask, Python, TypeScript, JavaScript, Node.js, Express, PostgreSQL, MongoDB, Auth0, REST APIs, GraphQL, OAuth 2.0, JWT.

A research assistant for the AIMING Lab at UNC-Chapel Hill, working under Dr. Huaxiu Yao's supervision.

Applied AI to biomedicine by training LLMs on medical text datasets with 500,000+ records and fine-tuning diffusion models to generate high-resolution medical images.

Improved model robustness against adversarial attacks, reducing error rates by 20% through training and fine-tuning.

Reduced diffusion models' computational costs by 35% while maintaining high-resolution image synthesis quality.

Developed retrieval-augmented generation (RAG) techniques, improving LLM knowledge retention and response accuracy by 40%.

Technology Stack: Python, PyTorch, LangChain, TensorFlow, Hugging Face Transformers, Diffusers, OpenAI API, Docker.

A UTA for COMP-126: Web Development for Everyone at UNC-Chapel Hill, under the guidance of Professor Tessa Joseph-Nicholas.

Provided academic support to 80+ students, both in-class and during office hours, improving their understanding of web development concepts.

Conducted 8–9 weekly office hours and review sessions, resulting in a 15% increase in student exam performance.

Graded and provided feedback on 200+ assignments and exams, maintaining a 100% on-time grading rate.

Communicated with the professor to enhance course materials, leading to a significant increase in student engagement and understanding of web development concepts.

Technology Stack: HTML, CSS, JavaScript, React, Node.js, Express, Git, GitHub, REST APIs, Bootstrap, Material-UI, Figma.

A learning assistant for COMP-210: Data Structures and Analysis at UNC-Chapel Hill.

Provided academic support to 150+ students, improving their understanding of data structures and algorithms.

Conducted 15+ weekly office hours and review sessions, resulting in a 20% increase in student exam performance.

Graded and provided feedback on 150+ assignments and exams, maintaining a 98% on-time grading rate.

Collaborated with professors to enhance course materials, leading to a 25% increase in student engagement in labs and assignments.

Mentored 10+ students on complex individual projects, helping them tackle advanced algorithmic challenges efficiently.

Technology Stack: Python, Java, Git, Data Structures (Arrays, Linked Lists, Trees, Hash Maps, Graphs).

Developed 20+ backend APIs and 10+ frontend features for FPT ICDP, FPT’s internal communication platform.

Enhanced the UI of ICDP using Figma, resulting in a 20% increase in user engagement and satisfaction.

Developed 5 new features of ICDP to streamline communication, document sharing, and task management, enhancing collaboration between FPT teams by 25%.

Contributed to NLP and data analysis projects for the chatbot feature of FPT's ICDP, leading to a 15% improvement in AI model fine-tuning and optimization.

Developed a new feature for FPT's ICDP that allows users to create and manage their own chatbots, increasing user engagement by 30%.

Technology Stack: Tensorflow, PyTorch, Optuna, React, Node.js, Express.js, MongoDB, PostgreSQL, Redis, Apache Kafka, RabbitMQ, Git, Kubernetes, Docker.

Developed the company’s full-stack English-version website and job application database that supports 50,000+ active users.

Led the entire project lifecycle from concept to deployment.

Created responsive, user-focused UIs, implementing a mobile-first design approach using Figma.

Increased operational efficiency by 40%, managing approximately 200 job applications monthly.

Deployed the system using Docker, AWS EC2, RDS, S3, Gateway, and Route 53, optimizing infrastructure costs by 20% and enhancing the company's global reach.

Helped establish a scalable infrastructure that accommodates future growth in user base and application submissions.

Technology Stack: PHP, React, Django, Figma, MongoDB, MySQL, Docker, AWS, Git.

Contributed to the design and development of vCloudcam’s security camera management website.

Enhanced the website’s UI/UX, boosting site performance by 30% and supporting user traffic of over 75,000 monthly visits.

Identified a performance issue and implemented a WASM-based solution that increased the video streaming system’s efficiency by 30%.

Rewrote the website’s backend code, improving the server’s response time by 40%.

Received the Best Intern Award for the 2023 Summer Internship Program by VNG Corporation.

Technology Stack: React, Angular, TailwindCSS, Webpack, Java, WebAssembly, Golang, Beego, Oracle Database, Red5, Docker, Git.

Collaborated with researchers at the Mandel School of Social Sciences on the Study of Human Rights in Southeast Asia and the Study of Southeast Asian Economics and Politics (SSEAP).

Assisted researchers with data entry, improved data processing efficiency by 30%, and executed 50+ data simulations and visualizations using Excel, Tableau, SAS, Plotly, and ggplot2.

Ensured data security and integrity for over 5,000 data points, reducing data errors by 25%.

Using Apache Spark and Hadoop, analyzed 10+ TB of data, identifying key trends and patterns that informed research findings and recommendations.

Reviewed and critically assessed 20+ research proposals, providing constructive feedback that enhanced the quality of research output by 40%.

Technology Stack: Microsoft Excel, Tableau, Hadoop, Spark, SAS, Python (Plotly), R (ggplot2).