Engineered collaboratively with various Departments at the University of Maryland, Built a robust and automated langchain pipeline to process research data via an ensemble of Amazon Nova Premier and Nova Lite models to ensure federal compliance of research proposals.
Stack: Streamlit, Python, MySQL, Langchain, AWS Bedrock, Hugging Face, AWS
Spearheaded the development of a dynamic Grain budgeting and data assessment tool, leveraging a comprehensive database to provide farmers with real-time insights into their crop budgets and financial planning.
Stack: NextJS, Radix UI, AWS Amplify, DynamoDB, AWS
Worked on the and
platforms,
to leverage zoominfo's best in class contact and buying signal data to provide customers with an end-to-end pipeline that can, source and hire a candidate or convert buyer research into actionable sales prospects!
Stack: AngularJS, NestJS, Groovy on Grails, Apache Solr, Jenkins, Google Cloud
Bringing real estate to the 21st century,
our vision is to change how real estate tranasctions are managed by title companies and real estate vendors.
Stack: NextJS, Express.js, MySQL, Github Actions, Vercel
Designed and backtested a proprietary, low-latency computational strategy for a weighted confidence model in C++ that synthesizes signals to generate high-conviction trading decisions using historical data from the Alpaca API.
The model achieved a backtested cumulative 2024 yearly return of 54.8% on $25K initial capital.
he model is deployed on a event-driven trading infrastructure on AWS, leveraging EventBridge to schedule Lambda functions that manage the lifecycle of a C++-optimized EC2 trading instance, minimizing operational costs.
stack: C++, Alpaca SDK, TA-lib, AWS
We use Hugging Face Inference Endpoints to deploy and scale a open source and custom LLM model to run inference through a openAI GPT based User interface.
Using open Web APIs, Voice input is also enabled based on the Web Speech API!
stack: NextJS, Langchain, Vercel AI SDK, Prompt Engineering