Headline

AI Engineer building agentic LLM systems | Multi-agent architectures, RAG pipelines, scalable AI workflows

About

Most AI projects stop at demos. The real challenge is building AI systems that can execute multi-step workflows reliably in production. I focus on designing agentic AI systems and LLM-driven pipelines that automate complex tasks. One example is ValuAI, where I built a multi-agent workflow for price intelligence. Agents handle data collection, processing, reasoning, and decision logic using a fine-tuned LLaMA model with QLoRA. During my research internship, I also developed a YOLOv8 computer vision pipeline for aerial asset detection, improving performance by 12 percent through dataset expansion and preprocessing. My approach is system-first engineering. Instead of relying entirely on large models, I design architectures where tools, structured logic, and models collaborate inside a workflow. This reduces cost, improves reliability, and allows AI systems to scale beyond prototypes. Core stack includes Python, PyTorch, Transformers, LangChain, LLaMA fine-tuning, computer vision pipelines, and deployment using FastAPI and Docker. I enjoy building systems at the intersection of LLMs, agent orchestration, and real-world automation. If you're building agentic AI systems or LLM-powered products, I’d be happy to connect and exchange ideas.

Location

New Delhi

AI stack

antigravity

claude code

cursor

langchain

mcp

n8n

Work experience

Growth Manager at NxtWave, Jul 2026 - Present · 3 mo, Hyderabad

Branch Coordinator at Student Mentorship Program, IIT Roorkee, Sep 2023 - May 2025 · 1 yr 9 mos

Web Developer at Students' Technical Council, Sep 2023 - Jun 2024 · 10 mos

Education

Bachelor's degree, Electronics and Communications Engineering, Indian Institute of Technology, Roorkee

Public links

github: https://github.com/Aryan-any