Looking for a job

Ayush Dadhich

@ayushdadhich

Message

I build production LLM serving, RAG, and MLOps systems in Python.

India

Message

What I'm looking for

I'm seeking remote or hybrid roles where I can build and operate production LLM inference, RAG, and end-to-end MLOps systems.

I've benchmarked vLLM, Ollama, and NVIDIA Triton for LLM serving, improving throughput by 145% and reducing time per output token by 67% through scheduler and parallelism tuning. In my independent engineering practice, I also built Argus, a self-corrective visual-document RAG system that improved correct answers from 15/22 to 19/22.

Previously at Tata Consultancy Services, I built Python ML pipelines across 150+ production servers to forecast capacity pressure and expand alert coverage. I also shipped Django and AWS applications at Snowstack and production mobile backends at Alkurn Technologies, bringing five years of experience across machine learning, software engineering, and production infrastructure.

Experience

Work history, roles, and key accomplishments

Current

Independent Machine Learning Engineer

Current

Self Employed

Aug 2024 - Present (2 years)

Independent work on LLM inference, retrieval systems, and production ML infrastructure. Benchmarked vLLM, Ollama, and NVIDIA Triton, and built a self-corrective RAG pipeline and an end-to-end MLOps project.

VLLM Ollama Triton Docker CUDA LangGraph Faiss PyTorch QLoRA DPO Streamlit fastAPI MLFlow GitHub Actions Prefect Google Cloud Run

Data Scientist

Tata Consultancy Services

Jan 2022 - Aug 2024 (2 years 7 months)

Built Python ML pipelines analyzing telemetry across 150+ servers for capacity forecasting and anomaly detection. Increased system alert coverage by 11% and established evaluation baselines for API gateways.

Python Pandas Matplotlib Time Series Analysis Anomaly Detection Machine Learning API Gateway