Vraj Patel
Data Science & Quantitative Finance
M.S. Data Science student at Stony Brook University. I build low-latency financial analytics engines,distributed ML pipelines, and scalable data-driven applications. Experienced in optimizing WebSocket streams and implementing real-time options strategies.
Technical Skills
Python
JavaScript
TypeScript
React
Next.js
FastAPI
Node.js
Tailwind
AWS
GCP
Docker
Kubernetes
Linux
Git
PostgreSQL
MySQL
MongoDB
Redis
TensorFlow
PyTorch
scikit-learn
Pandas
NumPy
R
Spark
Kafka
Airflow
Java
Scala
Streamlit
Plotly
Tableau
GraphQL
HTML5
CSS3
Python
JavaScript
TypeScript
React
Next.js
FastAPI
Node.js
Tailwind
AWS
GCP
Docker
Kubernetes
Linux
Git
PostgreSQL
MySQL
MongoDB
Redis
TensorFlow
PyTorch
scikit-learn
Pandas
NumPy
R
Spark
Kafka
Airflow
Java
Scala
Streamlit
Plotly
Tableau
GraphQL
HTML5
CSS3
Python
JavaScript
TypeScript
React
Next.js
FastAPI
Node.js
Tailwind
AWS
GCP
Docker
Kubernetes
Linux
Git
PostgreSQL
MySQL
MongoDB
Redis
TensorFlow
PyTorch
scikit-learn
Pandas
NumPy
R
Spark
Kafka
Airflow
Java
Scala
Streamlit
Plotly
Tableau
GraphQL
HTML5
CSS3
Python
JavaScript
TypeScript
React
Next.js
FastAPI
Node.js
Tailwind
AWS
GCP
Docker
Kubernetes
Linux
Git
PostgreSQL
MySQL
MongoDB
Redis
TensorFlow
PyTorch
scikit-learn
Pandas
NumPy
R
Spark
Kafka
Airflow
Java
Scala
Streamlit
Plotly
Tableau
GraphQL
HTML5
CSS3
Featured Projects
SBU VibeCheck – Real-Time Campus Event Platform
OCT 2025
- Built full-stack event aggregation platform in 36 hours (Hackathon Winner), implementing WebSocket-based real-time synchronization across clients with sub-200ms latency.
- Architected OAuth 2.0 dual-token flow (Auth0 for .edu verification) with server-minted anonymous JWTs for privacy-preserving sessions.
- Integrated Gemini 1.5 Pro with parallel function calling, parsing natural language to SQL queries via custom tool schemas, reducing query formation time by 85%.
- Designed stateful sentiment reporting system, broadcasting delta updates via Redis Pub/Sub to maintain real-time consistency.
- Deployed React SPA on Cloudflare Pages and FastAPI backend on GCP Cloud Run (auto-scaling), implementing TLS and CORS policies.
ReactFastAPIPostgreSQLWebSocketsRedisGoogle GeminiGCP Cloud Run
VociTrade - Voice-Activated Stock Trading
Sep 2026
- Engineered a high-performance backend using FastAPI and WebSockets, implementing a custom binary stream decoder to process tick-by-tick market data with sub-millisecond latency.
- Developed algorithmic execution logic for dynamic position sizing (e.g., % NAV allocation) and automated risk management protocols like global square-offs and margin validation.
- Integrated Large Language Models (Gemini) to translate unstructured natural language into structured financial transactions, utilizing RAG for context-aware market sentiment analysis.
- Built a floating "Heads-Up Display" (HUD) using Electron and React that overlays trading portals, enabling hands-free discretionary trading via custom wake-word activation.
PythonFastAPIDhan APISpeechRecognition.js
Work Experience
FinTech Data Scientist Intern
Flits (India)
Feb 2025 – May 2025- Built real-time options analytics engine processing NSE/BSE tick data, implementing asyncio WebSocket client with 40% reduction in monitoring overhead.
- Engineered ETL pipeline parsing 170K-row instrument master file, applying SIC code mapping and outputting normalized PostgreSQL schema.
- Developed sector rotation tracker correlating NIFTY movements, implementing z-score normalization to identify top decile outperformers.
View Full Diagram
📷 Real-time Options Analytics Engine Architecture
Data Analyst Intern
YHONK – Noise Pollution Mitigation
May 2024 – Jun 2024- Built distributed web scraping pipeline (Selenium Grid) with rotating proxies, extracting 50K+ school records with exponential backoff retry logic.
- Executed geospatial analysis on 2.6M GPS-tagged honking events, implementing PostGIS queries and R-tree indexing to compute violation hotspots.
- Applied ARIMA time-series decomposition on hourly violation counts, detecting 171% weekday anomaly and 130% seasonality.
View Full Diagram
📷 Geospatial Analysis of 2.6M Honking Events
International Research Intern
AIT Brain Lab (Thailand)
Jul 2023- Fine-tuned T5-base and FactorSum (BART-based) transformers on ParaSCI dataset, achieving 0.42 ROUGE-L score for abstractive summarization.
- Built Flask REST API with Celery task queue for async inference, implementing request batching to reduce P95 latency from 4.2s to 1.8s.
Education
Stony Brook University
M.S. in Data Science
New York, USA • 2025 – 2027
Pandit Deendayal Energy University
B.Tech in Computer Engineering
India • 2021 – 2025
CGPA: 8.72
Certifications
AWS Solutions Architect (Dec 2025)Alteryx Designer Core (2027)NLP Program, AIT (2023)
Get In Touch
Have a question or want to work together? Drop me a message and I'll get back to you as soon as possible.
Send me a message