Overview
I design and build production AI systems end-to-end: LangGraph orchestration pipelines , RAG architectures with fuzzy entity resolution, WebLLM on-device inference over WebGPU/WASM, and structured state extraction using JSON-mode reasoning models grounded in persistent vector stores.
My AI work is backed by 9 years of cloud engineering on AWS : serverless architectures, cross-region active-active infrastructure, GraphQL and REST APIs, and NoSQL databases at billions-of-records scale . I build AI systems that are integrated into real cloud infrastructure, not isolated experiments.
As a Solutions Architect I have a consistent track record of taking stakeholder ideas from concept to shipped product without hand-holding: translating ambiguous requirements into concrete system designs, owning the architecture decisions, and delivering across the full stack independently.
Certifications / Awards




Skills | Experience | Level |
|---|---|---|
LangChain | 2 years | Proficient |
RAG | 2 years | Proficient |
Stable Diffusion | 2 years | Skilled |
LLM Templating | 2 years | Proficient |
| 14 years | Expert |
TypeScript | 7 years | Expert |
| 5 years | Advanced |
C++ | 4 years | Skilled |
Vulkan | 1 year | Skilled |
| 3 years | Advanced |
| 3 years | Proficient |
| 8 years | Advanced |
| 11 years | Expert |
| 5 years | Expert |
| 2 years | Proficient |
| 3 years | Advanced |
Flask | 2 years | Skilled |
| 2 years | Skilled |
Relevant Projects
Money Studio
Founder & Principal Engineer
Software Startup
Built as founder and principal engineer : a TypeScript monorepo shipping a financial visualization platform to 5 platforms (web PWA, iOS, Android, Electron, Tauri) from a single codebase
On-device AI assistant powered by WebLLM (WebGPU/WASM) with a custom RAG pipeline grounded in the user's financial data via an encrypted local vector store
Architected a knowledge graph data layer, a node-based visual logic editor , and an AWS CDK backend with full IaC security scanning ( SAST , SCA , DAST , Checkov, OpenSSF Scorecard)
Integrated 5 Web3 wallet connectors (MetaMask, Coinbase, Phantom, Binance, Gnosis Safe) to include live crypto holdings in the financial knowledge graph
Interactive Narrative World Engine
Local LLM Simulation System
Personal Project
Designed an 8-node LangGraph pipeline for AI-assisted fiction: pre-response schedule reconciliation and spatial analysis ground every model response in consistent world state before generation
Fuzzy entity resolution via persistent Chroma vector stores prevents hallucinated inventory mutations; a secondary reasoning model (JSON-only, lower temperature) writes all state deltas deterministically to SQLite
Scene-aware image generation pipeline reads SQLite world state to compose diffusion prompts, select LoRAs , and choose sampler parameters entirely from structured data
QuantPy
Python Portfolio Optimization Library
Personal Project
Python library implementing Modern Portfolio Theory : Sharpe ratio maximization, minimum variance optimization, and efficient frontier tracing via SciPy SLSQP constrained optimization
Implemented the closed-form Two-Fund Separation Theorem for exact analytical weight solutions alongside the numerical optimizer, enabling comparison of both approaches on the same frontier
Rendered interactive Plotly efficient frontier charts from a universe of 40+ tickers, overlaying cumulative return traces for equal-weight, Sharpe-optimal, and benchmark portfolios (SPY, QQQ)
Employment History
Senior Software Engineer
Bluebook (Digital SAT) by The College Board
(4 yrs 7 mos) Aug. 2021–March 2026
Maintained and added features to a cloud service providing a centralized blockchain ledger for each student to detect cheating
Implemented observability as code and monitored the health of cloud resources including Lambda Invocations , DynamoDB Throttles , and API Gateway Errors
Maintained three massive NoSQL databases of millions of student exam responses, each containing billions of records
Budgeted and administered Stress Tests on cloud infrastructure and resolved performance bottlenecks
Removed 31 critical vulnerabilities in client dependencies such as Prototype Pollution and Template Injection
Refactored the React PWA core lifecycle to optimize CPU usage on millions of devices
Benchmarked compression libraries like pako , uzip , and fflate to choose and implement compression for our API POST payloads
Rebuilt the client build pipeline from Webpack to Vite , reducing build time by over 80%
Implemented dozens of UI updates to the Bluebook client to deliver new features and improve UX
Implemented hundreds of Unit and Integration tests using Jest and Playwright to reach 100% test coverage on multiple subsystems
Implemented JScrambler to obfuscate the client Javascript bundle while also minimizing the cost of obfuscation and bundle size
Upgraded the client build pipeline to upload sourcemaps to JScrambler and Sentry in production, preserving obfuscation while allowing developers to debug production
Oversaw the migration of AWS serverless cloud infrastructure from one region to use cross-region redundancy to improve availability and reduce latency
Analyzed and optimized the client Javascript build process to remove duplicates and unecessary libraries to reduce client bundle size by over 40%
Implemented chunking and exponential backoff to client API calls to manage server load and avoid rate limiting
Contributed to the architectural patterns integrating the cross-platform PWA client with the native OS wrappers
Contributed in the architecture and development of productizing the source code by extracting content into a CMS
Senior Full-Stack Engineer
100k Mentor Challenge by ProMazo, Inc.
(1 yr 4 mos) May 2020–Sept. 2021
Contributed over 220k additions and 140k deletions (lines of code) on a React Native codebase
Trained more than 15 developers on how to contribute to the codebase efficiently and problem-solve within the technology stack
Implemented more than 20 pages of a cross-platform app using React Native and Expo
Implemented an app-wide navigation system on the frontend using React Navigation & Redux
Managed the CI/CD Pipeline of the React Native app to the iOS / Android app stores
Managed the CI/CD Pipeline of all deployed environments of an AWS Cloud Infrastructure using AWS Amplify
Implemented quality-assurance and living-documentation tools including Jest, Storybook, Prettier, and Docsify
Solutions Architect & Senior Full-Stack Engineer
Mimino by ProMazo, Inc.
(2 yrs 2 mos) July 2019–Sept. 2021
Contributed over 270k additions and 243k deletions (lines of code) on a React PWA codebase
Modeled a GraphQL schema that has grown to 60+ models and 150+ connections and is still scaling
Implemented the mentioned GraphQL schema as an API in AWS using AppSync, CloudFormation, Amplify, and DynamoDB
Built an internal CLI tool to automate boilerplate code generation and improve developer efficiency by ~30%
Designed and implemented a scalable Aspect-Oriented frontend architecture to support complex data manipulations with React, Redux, Redux-Thunk, and the Amplify SDK
Software Architect Intern
Developer Tools by Looker Data Sciences, Inc.
(2 mos) June 2019–Aug. 2019
Used Node.js with TypeScript to build a developer CLI tool that shows the differences in Webpack bundles to track frontend payload size changes due to the code changes in a given PR
Integrated developer CLI tool into the existing CI/CD stack by working with GitHub, Jenkins, and AWS S3
Developed dozens of integration tests using Puppeteer and an in-house tool built on top of Cucumber
Cloud Backend Engineer
Viking by The Bank of New York Mellon Corp.
(1 yr 3 mos) Sept. 2018–Dec. 2019
Designed a SQL database with 8 tables and implemented via AWS RDS with encryption and disaster-recovery backups
Created a REST API using API Gateway and connected to the RDS databases via AWS Lambda serverless functions
Used AWS Cognito to implement a federated authentication system that conforms to the bank's strict security standards
Built a big-data processing system using AWS Lambda, Kinesis, CloudWatch, S3, and Step Functions to ingest financial data
Software Engineer Intern
Admin Tools by Looker Data Sciences, Inc.
(2 mos) June 2018–Aug. 2018
Used Javascript and Node.js to build a Chrome Extension for internal developers
Configured and deployed Node.js REST API to handle requests on Heroku
Integrated a developer CLI tool into the existing CI/CD stack by working with GitHub, Jenkins, and AWS S3
Developed dozens of integration tests using Puppeteer and an in-house tool built on top of Cucumber
Full Stack Engineer
LASSO by University of Rochester IT Department
(2 yrs 3 mos) Feb. 2017–May 2019
Maintained and extended the legacy Django server-side-rendered MVC codebase in Python , resolving bugs and shipping new features on LASSO , the in-house employment and inventory management system
Joined a team of 5 engineers to rewrite LASSO from scratch in a separate repository , replacing a codebase tangled with cross-cutting concerns
Delivered a full platform rewrite in 4 months
Built a REST API on the backend with Express.js and Node.js
Implemented frontend logic with Vue.js for the new LASSO platform
Digital Media Specialist
Rettner Hall Digital Media Lab at University of Rochester
(2 yrs 7 mos) Oct. 2016–May 2019
Served as the primary customer-facing representative of the lab, greeting and assisting a high volume of students daily with a focus on hospitality and satisfaction
Provided technical support to college students using digital media software including Photoshop , Illustrator , Blender , Unreal Engine , Unity , and more
Resolved student concerns and equipment issues on the spot, maintaining a calm, professional demeanor under a busy, fast-paced environment
Managed and maintained the lab's virtual reality equipment, 3D printers , and A/V equipment
Bank Teller
First National Bank of Louisburg
(1 yr 1 mo) June 2015–July 2016
Followed legal procedures for the proper processing of checks , cash , and loans , and assisted in the management of customer accounts
Helped customers troubleshoot and resolve financial discrepancies such as searching for missing funds and settling balances
Accurately handled over $50,000 in cash each day
