Hi, I’m Jashwanth S 👋 — an AI/ML Engineer focused on building production-ready GenAI systems, RAG pipelines, and scalable data platforms. I enjoy turning complex AI ideas into reliable, cost-efficient systems that work in the real world.
🔭 Currently working on: AI automation workflows, OCR & document intelligence, and large-scale data pipelines
👯 Looking to collaborate on: Open-source AI/ML, GenAI apps, LLM-powered tools, data engineering, and automation (Python, Docker, n8n)
🤝 Looking for help with: Advanced LLM fine-tuning, production-grade MLOps, agent orchestration, and scalable GenAI systems
🌱 Currently learning: LLM internals, AI agents, cloud-native ML, MLOps best practices, and low-latency inference
💬 Ask me about: Machine Learning, Deep Learning, Generative AI, RAG systems, OCR, data engineering, MLOps, and system design
⚡ Fun fact: I enjoy debugging as much as building—most of my learning comes from breaking things and fixing them 🚀
- 🔹 LLM-powered RAG systems with evaluation (BLEU, ROUGE, BLEURT)
- 🔹 OCR & document intelligence pipelines for structured data extraction
- 🔹 Scalable data pipelines using Airflow, Spark, and SQL
- 🔹 Real-time ML inference APIs (FastAPI, Flask)
- 🔹 GenAI automation using agents and workflow orchestration
| Project | Description | Tech Stack |
|---|---|---|
| SpeakEasy 🗣️ | Real-time Indian language translation system achieving 23.4 BLEU (En-Hi). Supports 6+ languages with <1.8s latency. | Transformers HuggingFace Python |
| Enterprise RAG 🤖 | Retrieval system processing 5,000+ docs with 2.3s retrieval time and 85% relevance accuracy. | LangChain FAISS Streamlit |
| IoT Predictive Maint. ⚙️ | Forecasting device failures with 92% accuracy and 87% recall, deployed on AWS. | AWS XGBoost Flask |
| Waste Mgmt System ♻️ | CNN-based waste detection with 92% accuracy (Patent Published: 202341082049 A). | OpenCV TensorFlow CNN |
Always learning. Always building. Always improving.