Data Science & Deep Learning background | RAG Workflows | Context Engineering | LangFuse & LangSmith Expert
I’m Ayush Sonu, an AI Engineer with a background in Generative AI, Data Science and Deep Learning.
I focus on building agentic AI workflows, context-aware RAG systems, and evaluation-driven pipelines to make LLMs more accurate and reliable.
- 💼 AI Engineer at Impressico Business Solutions
- ⚙️ Experienced with LangChain, LangGraph, LangFuse, LangSmith, Qdrant, and FAISS
- 🧩 Skilled in retrieval-augmented generation, context optimization, and evaluation frameworks
Retrieval-Augmented Generation • Prompt Engineering • Dynamic Context Flow
Knowledge Routing • Evaluation-based Optimization
Python • LangChain • LangGraph • OpenAI • Hugging Face
Django • FastAPI • Celery • AWS • Redis • boto3
LangFuse • LangSmith • Ragas • TruLens • Prometheus Eval
AI systems that reason, adapt, and persist — beyond single responses.
I design:
- 🌐 RAG systems with adaptive retrieval
- 🧩 Agentic workflows powered by LangGraph
- ⚙️ Reliable pipelines with observability and feedback loops
- 📊 Evaluation frameworks using LangFuse, LangSmith, and Ragas
I enjoy mentoring others:
- 🚀 How to start with LangChain / LangGraph
- 🧠 Building RAG pipelines from scratch
- 🧩 Designing reliable multi-agent systems
- 🎯 Using LangFuse / LangSmith for evaluation & monitoring
Helping developers move from prompting → to engineering intelligent systems.
- Contributed to Aegra — an open-source framework implementing the LangGraph Agent Protocol.
Focused on improving agent hosting on your own infrastructure reliabily at scale.
