|
identity:
username: ayushsyntax
role: Data Scientist & Machine Learning Engineer
education: B.Tech in Artificial Intelligence & Data Science
current_mode: Building reliable AI products, not just demos
system:
focus: [GenAI, RAG, Agentic Systems, MLOps, Deep Learning]
stack: [Python, SQL, TensorFlow, PyTorch, XGBoost, LangChain, LangGraph]
infra: [Docker, MLflow, DVC, AWS ECS, GitHub Actions, FastAPI]
data: [Pandas, NumPy, Matplotlib, Seaborn, ChromaDB]
workflow: [Experiment, Evaluate, Trace, Optimize, Deploy]
currently_building:
- Reasona: self-improving retrieval and reasoning engine
- Agentic research workflows with memory and tool routing
- Deployment-first ML systems with observability
philosophy:
- elegant models
- reproducible pipelines
- measurable outcomes |
$ whoami
ayushsyntax
$ mission --current
Build AI systems that are accurate, observable, and production-ready.
$ values --core
reasoning > hype
shipping > showcasing
clarity > complexity|
Production-oriented medical imaging pipeline using EfficientNetV2-S, MLflow, DVC, Docker, FastAPI, and AWS ECS. |
Leakage-safe regression architecture with Optuna tuning, FastAPI inference, Streamlit UI, and AWS deployment. |
|
LangGraph-based multi-tool agent with ChromaDB memory, Groq inference, SQLite checkpointing, and LangSmith tracing. |
Self-improving RAG engine inspired by iterative retrieval and correction loops for grounded answers. |
