Projects
A showcase of my engineering projects, demonstrating practical application of skills and innovative problem-solving.
Multi-Tenant AI-Driven SaaS Platform
This is an AI-driven SaaS platform that supports the full RFP and RFQ lifecycle, including requirement analysis, question and answer management, proposal authoring, compliance measurement, and team evaluation. By leveraging contextual AI and historical proposal intelligence, it significantly reduces response time while enabling organizations to produce consistent, compliant, and highly competitive submissions at scale.
Production Medical Imaging Pipeline
This platform is an AI-powered computer vision system that automates the labeling, classification, and grading of echocardiogram images, delivering highly accurate, real-time diagnostic insights. By replacing time-intensive manual analysis with scalable medical imaging AI, the platform dramatically reduces analysis time while improving diagnostic precision and clinician workflow efficiency.
AI Performance Monitoring Chatbot
This system is a production-grade conversational system that enables real-time monitoring, analysis, and troubleshooting of AI models and enterprise services through a unified natural-language interface. It automates large-scale data retrieval across logs, metrics, and alerts, significantly reducing incident resolution time and accelerating support workflows in high-throughput environments.
AI Construction Compliance Platform
This is an AI-driven automation platform designed for a seismic construction company to generate project documentation and regulatory compliance reports with greater accuracy and consistency. The platform features React-based dashboards integrated with Elasticsearch, providing enhanced reporting visibility, observability, and operational insights.
Predictive ML Model Serving
This is a predictive machine learning platform built to generate real-time business insights for a SaaS product and surface them directly within core user workflows. The system integrates production ML inference into MERN-based dashboards, enabling data-driven decision-making through embedded predictions and analytics.
Geospatial & Climate Analytics
This is a scalable AI platform designed to support large-scale geospatial data processing and real-time predictive analytics for climate and energy use cases. The system integrates distributed ETL pipelines, ML inference services, and high-performance APIs to deliver responsive geospatial insights to frontend dashboards under cost-efficient cloud constraints.
Fine-Tuning Transformer for NLP
Fine-tuned pre-trained transformer models for multiple NLP tasks using PyTorch and early Hugging Face Transformers, optimizing training stability and accuracy under limited compute constraints. Applied techniques such as learning-rate scheduling, layer freezing and unfreezing, gradient accumulation, and batch-size tuning to improve convergence and reduce inference cost.
Semantic Search & Recommendation
This is a semantic search and recommendation engine built to improve discovery and reuse of information across large collections of unstructured text. The system leverages transformer-based embeddings and vector similarity search to deliver fast, relevant results that support operational efficiency and decision-making.