Sunnyvale, California, United States
🙋♂️Me: - Experienced Engineer: Proficient in backend development and machine learning model training - Multi-disciplinary: Holds majors in Finance, Business, Management, and Engineering - Fast Learner: Capable of quickly mastering new techniques 🧑💻Skills: - Languages: Python, JavaScript, C#, Java, C, C++, F# - Technologies: AWS, Azure, Docker, Git, ASP.NET Core, Node.js, GDB, GraphQL - Databases: MongoDB, ChromaDB, MS SQL Server, MySQL
• Built and scaled mission-critical backend services powering Echo Show’s LLM-driven multimodal experiences, sustaining 100K+ TPS with high availability across global traffic. • Led the architecture and delivery of a multi-agent LLM orchestration platform with RAG and hybrid retrieval (BM25 + vector search), significantly reducing onboarding friction and operational overhead across teams. • Architected an LLM-driven conversational system powering Alexa+ AI experiences across Echo Show, Fire TV, mobile, and web platforms, enabling real-time, context-aware multimodal interactions through scalable data pipelines, low-latency model serving infrastructure, and continuous fine-tuning. • Built scalable data ingestion and processing pipelines to collect, transform, and manage large-scale ground truth datasets for AI evaluation and continuous model improvement workflows.
Course: 18-631 Introduction to Information Security
Topic: Creating an Ecosystem of Distributed Systems - Raft DB Sharding Advisor: Leonardo da Silva Sousa • Designed and implemented a data sharding proxy server utilizing the Consistent Hashing algorithm with Java Spring Boot, achieving a 40% enhancement in query performance and facilitating a 50% increase in system scalability • Optimized Raft algorithm and implemented the Two-Phase Commit (2PC) protocol in C++ to ensure data consistency • Refactored the gRPC-based metrics collector for gathering data from both RaftDB and 2PCDB cluster
• Developed a scalable, cloud-native architecture for transitioning a Digital Signal Processors (DSP) hardware solution to AWS. • Achieved a 50% cost reduction and 3x processing speed by implementing a multi-threaded TCP server in Python • Developed an automated AWS SDK platform using Python boto3, Lambda, and Docker, achieving a 10 times reduction in deployment time and significantly enhancing usability
• Developed a Large-Language Model (LLM)-based data analysis web application using Python FastAPI, featuring automated code generation, execution, and report generation, which reduced data scientists' development time by 70% • Implemented agent switch mechanisms employing the strategy pattern, coupled with knowledge vector database persistence through ChromaDB, resulted in a 60% reduction in operational costs • Pioneered a Docker-integrated, pluggable storage microservice with a GraphQL-based unifying interface, using JavaScript, Apollo Server, and MongoDB, reducing cost overhead by 80% while supporting 500 queries per second (qps) • Engineered an LLM chatbot featuring Retrieval-Augmented Generation (RAG), indexing 10,000+ documents, enhancing retrieval accuracy by 40% and enabled real-time data inquiries and responses for over 5,000 users across 6 business units • Developed a Power BI report posting platform, streamlining report publishing across global branches • Integrated Power BI SDK into a .NET Core web application, reducing IT workloads by 50% and license costs by 80% • Engineered RESTful APIs for embedded report functionality, boosting internal usage rates within the firm to 50%
• Handled design and development of NLP (Market Intelligence Platform) projects • Enhanced a Named Entity Recognition (NER) model using a BERT-based transformer and a crawling feedback loop, increasing accuracy from 55% to 70% • Collaborated across teams to integrate millions of data points using SQL on the Denodo platform and digitalized over 10 dashboards with Tableau
• Designed, developed and implemented MI (Market Intelligence) platform. • Developed and implemented web crawlers and keyword dictionaries in Python to predict Supply Chain trends, reducing decision-making time and successfully forecasting the potential material shortage resulting from the 2021 China outage • Developed and deployed over 4 Web Crawlers to a task scheduler on a Windows Server using Python Requests and Selenium, storing data in SQL Server