Project: Instagram Clone — 5-Database Backend Architecture
Engineered a high-performance backend for an Instagram-style application, splitting different types of data across 5 distinct databases to handle millions of fast user interactions with low latency.
Key Technical Implementations:
• Cassandra (Feeds & Timelines): Handled massive read/write traffic by partitioning user feeds chronologically to bypass slow database joins.
• Redis (Real-Time State): Utilized atomic increment operations to handle millions of likes and instant UI feedback without race conditions.
• Neo4j (Social Graph): Built a graph model to map user relationships and power fast follower recommendation queries.
• MongoDB (User & Comment Metadata): Managed semi-structured, flexible user profiles and nested, threaded comment replies.
• MinIO (Media Storage): Configured an Amazon S3-compatible object bucket system to store raw video reels and images cleanly.
• Docker Compose: Containerized the entire microservices architecture to allow each component to scale independently.
1
9
Project: Confidex Analytics — Computer Vision-Driven Behavioral Scoring Pipeline
Developed a high-accuracy, research-backed computer vision and deep learning pipeline designed to analyze non-verbal upper-body ergonomics and hand kinematics from video streams to calculate objective confidence metrics.
Key Technical Implementations:
• Landmark Feature Extraction: Built a real-time frame processing pipeline using OpenCV and MediaPipe to convert raw interview videos into structured sequential tracking data, capturing spatial coordinates of physical gestures and skeletal movements from 200+ video datasets.
• Deep Learning Architecture: Evaluated and compared recurrent neural networks (LSTM, GRU, BiLSTM) to model temporal gesture development across frame streams, selecting and optimizing a high-performance BiLSTM sequence model that achieved a verified 82.79% prediction accuracy.
• Temporal Modeling & Softmax Aggregation: Preserved the model's entire uncertainty distribution layer by retaining complete softmax outputs. Structured a temporal pipeline that aggregates these class probabilities across strict 1-second rolling window frames to maintain maximum representation stability.
• Research-Based Weighted Fusion: Engineered a quantitative scoring engine that maps normalized gesture frequencies to research-derived behavioral coefficients (Hands on table: 0.322, Hidden hands: 0.264, Gestures on table: 0.207, Self-touch: 0.207). Programmed dynamic inversions for negative behavior inputs (Score = 10 - Negative) to output a clean, unified confidence rating out of 10.
• Backend Deployment: Exposed the finalized end-to-end deep learning scoring pipeline via a high-performance FastAPI web service layer for production readiness.
0
10
Flight Delay Prediction Web App
0
15
Engineered a scalable end-to-end cloud data pipeline processing 356,000 daily trip records via Auto Loader on AWS S3 buckets. Architected a robust Medallion (Bronze/Silver/Gold) storage environment using Unity Catalog to orchestrate ingestion, data quality validation checks, and optimized analytics materialized views.
1
29