I am a Development Team Lead and AI/Backend Engineer focused on building production-grade AI systems, scalable Python services, and real-world automation platforms.
- Leading 20+ Engineers at BuiltPulse and Nexxaura across backend, AI orchestration, and delivery workflows.
- Building agentic AI systems, multi-agent pipelines, and RAG applications with LangGraph, LangChain, OpenAI, Anthropic, and vector databases.
- Designing scalable backends with Django REST Framework, FastAPI, Celery, RabbitMQ, Redis, PostgreSQL, Docker, AWS EC2/S3, and CI/CD.
- Experienced in medical billing automation, ERP platforms, bioinformatics, NLP, computer vision, and document-grounded AI assistants.
- Published research in MDPI on transformer-based NLP and deep learning for computer vision.
| Role | Organization | Focus |
|---|---|---|
| Development Team Lead | BuiltPulse | Enterprise ERP, agentic workflows, multi-tenant backend, real-time collaboration |
| Development Team Lead | Nexxaura | AI medical billing CRM, RCM automation, LangGraph orchestration |
| AI/Backend Engineer | Independent Projects | RAG assistants, NLP systems, ML platforms, automation tools |
| Project | Tech Stack | What I Built |
|---|---|---|
| CIA - Enterprise ERP Platform | Django, DRF, LangGraph, WebRTC, Celery, Redis, Docker, AWS, Nginx | Multi-tenant ERP with RBAC, HR, Finance, Project Management, Compliance, AI task suggestions, real-time collaboration, CI/CD, Sentry, and Grafana. |
| MedSynthea - AI Medical Billing CRM | LangGraph, LangChain, Django, FastAPI, Celery, Redis, Docker, AWS | AI-driven CRM for eligibility verification, insurance verification, medical coding, denial prediction, claim scrubbing, and EOB processing. |
| PON-P3 Pathogenicity Prediction System | Django, Celery, RabbitMQ, Pandas, ReportLab | Bioinformatics web app for Lund University with async batch processing, genomic data handling, automated PDF reports, and email delivery. |
| AI-Generated Cryptocurrency Tweet Detection | PyTorch, Transformers, Ollama, NLP | Transformer classifier achieving 99% accuracy for detecting AI-generated financial content. Published in MDPI 2025. |
| Multi-Class Visual Cyberbullying Detection | TensorFlow, Keras, Computer Vision | Deep learning pipeline reaching 98% accuracy for visual cyberbullying detection. Published in MDPI 2024. |
| AccountingBot - RAG Assistant | LangChain, Selenium, Pinecone | Domain-specific RAG assistant with web scraping, vector search, conversational memory, authentication, and session management. |
| EduAI - AI Educational Platform | Groq, LangChain, Llama Vision, RAG, OCR | Educational AI platform with document Q&A, video interaction, quiz generation, Word export, and OCR. |
- Human or AI? Transformer-Based Detection of AI-Generated Financial Content - MDPI, 2025
- Multi-Class Visual Cyberbullying Detection using Deep Neural Networks (CVID Dataset) - MDPI, 2024
- Associate Data Scientist - DataCamp
- Data Analyst - DataCamp
- Machine Learning Specialization - DeepLearning.AI / Stanford University
- Python for Everybody - University of Michigan
Building AI systems that move from prototype to production.