AYUSHI KUKREJA

AI/ML Engineer | Data Scientist | NLP Specialist
Noida, IN.

About

Highly accomplished AI/ML Engineer with a robust background in developing and deploying advanced machine learning and natural language processing solutions. Proven expertise in building Retrieval-Augmented Generation (RAG) pipelines, knowledge graphs, and data-driven systems that deliver measurable impact. Adept at leveraging a diverse tech stack including Python, TensorFlow, LangChain, and various databases to solve complex problems and drive innovation. Seeking to apply leadership and technical acumen to challenging AI/ML roles.

Work

Humanitics Dimensions Pvt. Ltd.
|

AI/ML Engineer

Summary

Driving innovation in legal tech through advanced AI/ML and NLP solutions, focusing on enhancing document understanding, knowledge representation, and automation for legal processes.

Highlights

Engineered an advanced Retrieval-Augmented Generation (RAG) pipeline for legal Q&A, achieving 93% retriever accuracy by leveraging FAISS and adaptive reranking strategies.

Integrated Natural Language Processing (NLP) functionalities via OpenNyai pipeline, including Named Entity Recognition (NER), Rhetoric Roles analysis, and Judgment Summarization.

Implemented a scalable document summarization framework utilizing LangChain MapReduce Chain.

Designed and implemented a Neo4j-based knowledge graph for legal case management, optimizing querying with LangChain GraphCypherQA Chain for structured and full-text semantic search.

Automated court-mandated legal compensation calculations through MACT API integration.

Developed a multilingual document translation application for legal documents, fine-tuning YOLO 11 for layout detection and M2M100 418M for translation.

Chegg
|

Managed Network Expert (Freelance)

Summary

Provided expert-level academic support in computer science, delivering clear and accurate solutions to complex inquiries and aiding student comprehension of core concepts.

Highlights

Provided over 100 timely and precise solutions to complex computer science inquiries on the Chegg platform, demonstrating deep subject matter expertise.

Simplified complex programming, algorithms, and data structure concepts into clear, concise explanations for student comprehension.

Tanmay Sachin Foundation
|

Web Developer (Intern)

Summary

Contributed to enhancing the online presence and user engagement for a non-profit organization through front-end web development.

Highlights

Contributed to the development and maintenance of the foundation's website, enhancing user experience for over 50 daily unique visitors.

Optimized site compatibility across various devices, leading to a 20% increase in web traffic.

Education

University of Delhi

Bachelor of Science (Hons)

Computer Science

Grade: 8.2/10.00 CGPA

Agra Public School

High School Diploma

Secondary Education

Grade: 95.2%

Languages

English

Native

Skills

Programming Languages

Python, SQL, C++.

Data Analysis & Visualization

Pandas, NumPy, Matplotlib, Seaborn, Plotly, Power BI, Microsoft Excel.

Databases

MySQL, PostgreSQL, Neo4j (Graph Database).

Machine Learning

Scikit-Learn, TensorFlow, Keras, Natural Language Processing (NLP), Deep Learning, Computer Vision, Generative AI.

Statistical Analysis

Hypothesis Testing, Regression Analysis, Statistics, Problem Solving.

AI/ML Frameworks & Tools

PyTorch, TensorFlow, Scikit Learn, LangChain, LangGraph, Ollama, Hugging Face, CrewAI, OpenCV, TKinter, VLC.

Web Technologies

HTML, CSS, JavaScript, jQuery.

Projects

Text2SQL Retrieval Generation Agent

Summary

An advanced system designed to convert natural language queries into precise SQL commands, enhancing data accessibility across multiple database systems using large language models and semantic search.

Emotion based song Recommendation System

Summary

A real-time music recommendation system that analyzes user emotions to curate personalized playlists, integrating computer vision for emotion detection and a user-friendly interface.

Fraud Detection (Machine Learning)

Summary

A machine learning-powered system developed to identify and prevent fraudulent activities in online payment transactions by analyzing historical data and detecting abnormal patterns.