Looking for a job

Ayush Agrawal

@ayushagrawal5

Message

AI/ML engineer building RAG pipelines at Tenneco and generative AI workflows, including an ad poster generator.

India

Message

At Tenneco, I developed RAG pipelines combining LLMs, semantic search, and vector embeddings for information retrieval. I also built and evaluated hybrid machine learning models for data analysis and predictive tasks.

At Sunfinity Technology Solutions, I built AutoCleanML, an automated data-cleaning framework that uses rule-based and ML techniques to address missing values, outliers, and schema inconsistencies. I integrated LLM-assisted reasoning and developed preprocessing pipelines and REST APIs using FastAPI.

On my Desi-Scribe project, I designed an AI workflow that takes product understanding through to advertisement poster generation using three generative AI modules. I optimized its backend orchestration and inference pipeline to generate marketing creatives in 3–6 seconds with multilingual support.

Experience

Work history, roles, and key accomplishments

Data Analyst Intern

Tenneco

May 2026 - Jul 2026 (2 months)

Developed RAG pipelines integrating LLMs, semantic search, and vector embeddings for intelligent information retrieval. Built and evaluated hybrid machine learning models for large-scale data analysis and predictive tasks, addressing imbalanced datasets with advanced techniques.

RAG LLM Semantic Search Embeddings Machine Learning Data Analysis Class Weighting Re Sampling Prediction Threshold Optimization

AI/ML Intern

Sunfinity Technology Solutions

Jan 2026 - Apr 2026 (3 months)

Built an automated data-cleaning framework to detect and resolve missing values, outliers, and schema inconsistencies using rule-based and ML techniques. Integrated LLM-assisted reasoning with adaptive preprocessing pipelines and developed scalable REST APIs using FastAPI with RLHF-based feedback.

Python fastAPI Machine Learning Data Cleaning LLM RLHF REST APIs Scalable Preprocessing Pipelines

Software Development and Data Science Intern

TEKSOFT Labs

May 2025 - Jun 2025 (1 month)

Architected end-to-end ML pipelines including preprocessing and evaluation for 100K+ records, focusing on model deployment and monitoring to ensure production reliability, boosting training efficiency by 20% and achieving 92% model accuracy. Performed EDA on 250K+ records to identify 3 revenue leakage factors.

scikit learn SeaBorn Pandas EDA Machine Learning Model Deployment Monitoring