Austin, Texas Metropolitan Area
Software engineer in Austin building high-impact solutions at the intersection of fintech, AI, and data. I design scalable backend infrastructure, cloud architectures, and ML models to solve complex problems and deliver reliable products.
- Engineered backend telemetry APIs and integrated Loki/Prometheus for real-time Kubernetes observability and log aggregation. - Built and iterated React/TypeScript interfaces with custom tooling to enhance monitoring workflows and frontend performance. - Orchestrated full-stack vulnerability remediation using SAST/DAST tools, resolving CVEs across Node.js services and client-side assets. - Ensured system reliability and engineered over 30 patches to existing tools for a general availability (GA) release significantly ahead of schedule.
Developing local AI-enhanced features such as LLM finetuning, RAG applications, OCR deployments, and AI security for TwoKey’s analytics and research applications for small enterprises. Particular focus on security and data confidentiality.
Teaching Assistant for multiple foundational Computer Science courses. 10.014 Computational Thinking for Design: Provided in-class and after-hours Python programming instruction and support, and contributed to lab curriculum and assessment. 10.020 Data Driven World: Taught data science concepts and tools, and supported students in hands-on data analysis projects employing Flask, NoSQL, and machine learning libraries like Sklearn. 50.001 Information Systems & Programming: Taught students OOP principles, Java abstraction, and networking concepts, and contributed to course development and delivery.
Worked alongside Prof. Malika Meghjani, BlueRobotics and 6 undergraduate students on developing computer vision algorithms for marine research robots. Tested and verified efficacy of computer vision models using BlueRobotics autonomous underwater vehicles (AUVs) in waters around Singapore. Conducted image cleaning and analysis for real time datasets used by the AUV.
• Developed ETL pipelines to transform, denoise, and engineer ETF and ticker data, improving regression model accuracy for predicting monthly returns. • Reduced RMSE by 40% and improved directionality by 50%, significantly boosting model performance. • Compared LSTM, LTCNN, LTC with GRU, CTRNN, and ARIMA models for recession and stock trend prediction, evaluating both machine learning and statistical approaches for classification and forecasting. • Conducted extensive model evaluations to identify the best methods for stock price trends and recession predictions.