website logo

Udyan Sachdev

AI/ML Engineer

AI/ML Engineer with 4+ years driving $2.4M+ business impact at Accenture, HPE, and Duke Health. Expert in building scalable ML pipelines and automated solutions that power real-time decision-making and transform complex data into measurable value.

Key Achievements

  • Generated $2.4M+ in measurable business value
  • Deployed 15+ production ML systems
  • Reduced processing time by up to 60%
📬 Get in Touch

🎓 Education

Duke University

Aug 2023 – May 2025

Masters in Data Science

The Australian National University

Aug 2019 – May 2021

B.E. (Hons) – Mechatronic Systems

Vellore Institute of Technology

Aug 2017 – May 2019

B.Tech – Electronics and Communication Engineering

💼 Experience

Duke Health August 2024 – May 2025

Generative AI Engineer

Impact: Transformed clinical data accessibility for 200+ users
  • Architected and deployed a breakthrough Conversational Data Intelligence System, revolutionizing how clinical teams interact with ICU and financial datasets through natural language queries
  • Engineered Model Context Protocol (MCP) integration with LLM-powered chat interface, enabling sophisticated multi-turn conversations with contextual awareness
  • Achieved 95%+ accuracy in natural-to-SQL translation, seamlessly bridging the gap between business questions and data insights
  • Delivered 60% reduction in time-to-insight while dramatically improving data accessibility across cross-functional teams
  • Empowered 200+ business users to self-serve analytics, eliminating bottlenecks and accelerating decision-making processes
Python LangChain OpenAI SQL Docker AWS
Hewlett Packard Enterprise (HPE) June 2024 – December 2024

Data Scientist

Impact: $1.2M annual revenue uplift and 8% reduction in customer churn
  • Engineered automated data pipelines using Apache Spark and Airflow, achieving 35% reduction in data latency for critical product configuration processes
  • Built and deployed hybrid recommendation engine leveraging Spark, Hive, and Python, driving 15% increase in cross-sell conversions and 8% reduction in customer churn
  • Generated estimated $1.2M annual revenue uplift through intelligent recommendation systems and optimized customer targeting
  • Integrated complex structured and semi-structured data sources from multiple e-commerce channels into centralized AWS S3 + Redshift warehouse
  • Transformed key business functions: Inventory Management, Dynamic Pricing, Personalized Customer Experience, and Demand Revenue Forecasting
Python Apache Spark Airflow AWS Redshift Hadoop
Accenture August 2021 – July 2023

Data Science Consultant

Impact: $1.25M combined cost savings and revenue generation, 40% reduction in financial data processing time, 30% increase in web traffic
  • Architected advanced ETL pipelines and data models for corporate financial report analysis using NLP and Python, achieving 40% reduction in processing time while maintaining data integrity
  • Engineered large-scale distributed pipeline for solar energy forecasting across India using PySpark and Hadoop, improving prediction accuracy by 18% and enabling optimized energy distribution
  • Generated estimated $500K annual cost savings through intelligent energy forecasting models that transformed grid management and resource allocation strategies
  • Built sophisticated Hybrid Recommendation System for Data Marketplace, driving 30% increase in website traffic and 25% boost in user engagement metrics
  • Delivered projected revenue boost of $750K through personalized recommendation algorithms that enhanced user experience and marketplace conversion rates
  • Implemented cutting-edge Computer Vision models for coal excavation progress tracking and estimation, improving operational efficiency by 12% across mining operations
  • Recognized with prestigious Techie Award for exceptional expertise in AI/ML tools and measurable business impact across multiple client engagements
Python PySpark Hadoop NLP TensorFlow Computer Vision SQL Apache Spark
Technology Healthcare Bigdata December 2020 – February 2021

Data Scientist Intern

Impact: 92% F1 score in patient risk prediction
  • Developed comprehensive feature pipeline for patient risk modeling in healthcare settings, achieving exceptional F1 score of 0.92 through advanced feature engineering
  • Leveraged Google BigQuery, TensorFlow, and SQL to process large-scale healthcare datasets and build predictive models for patient outcome assessment
  • Implemented robust data preprocessing and feature selection techniques that improved model performance and clinical decision-making capabilities
  • Collaborated with healthcare professionals to ensure model interpretability and clinical relevance in real-world medical applications
Python TensorFlow Google BigQuery SQL Pandas Scikit-learn
Rezo.AI November 2019 – January 2020

Data Scientist Intern - Speech Analytics

Impact: 85%+ accuracy in Hindi speech recognition
  • Pioneered Automatic Speech Recognition (ASR) research focusing on Indian Hindi language transcription, addressing unique linguistic challenges and dialectal variations
  • Achieved breakthrough accuracy of 85%+ for Hindi speech recognition using advanced machine learning algorithms and acoustic modeling techniques
  • Explored and implemented cutting-edge ML algorithms including Deep Neural Networks and Hidden Markov Models for speech pattern recognition
  • Contributed to expanding AI accessibility for Hindi-speaking populations through improved speech-to-text capabilities
  • Optimized model performance through feature engineering, data augmentation, and hyperparameter tuning for Indian language processing
Python TensorFlow Deep Learning NLP Speech Processing Acoustic Modeling

Featured Projects

AI Sales Agent

AI Sales Agent automates lead generation and outreach, enabling businesses to identify, qualify, and engage high-value prospects for maximum revenue impact.

AI Sales Agent

PlotVerseXR

Powered by OpenAI, PlotVerseXR converts natural language prompts into intelligent 3D data visualizations. It uses AI to generate custom Plotly code from user input, transforming datasets into immersive XR experiences viewable on Meta Quest—enabling intuitive exploration of complex data like LLM embeddings and clustering patterns.

PlotVerseXR

LSTM Stock Forecasting App

Created a Flask web app for stock trend forecasting using LSTM models, Azure Databricks, and Dockerizedmicroservices, improving prediction accuracy by 12%.

LSTM Stock Forecasting App

Advanced Visual Question Answering and Text-Image Retrieval

A robust Visual Question Answering (VQA) system by integrating the CLIP (Contrastive Language–Image Pre-training) model with a Retrieval-Augmented Generation (RAG) framework

Advanced Visual Question Answering and Text-Image Retrieval

SmartFood QA: Vision-Language System for Food Recognition and Interaction

Developed a multimodal app using Meta LLaMA for food recognition and Q&A, with a Streamlit UI, Rust-Python backend, and end-to-end AWS deployment via Docker and GitHub Actions.

 SmartFood QA: Vision-Language System for Food Recognition and Interaction

NutriVision: AI-Powered Dietary Tracking from Food Images

Built an AI app that identifies food items and estimates nutrition using image recognition and generative AI, streamlining dietary tracking.

NutriVision: AI-Powered Dietary Tracking from Food Images

Databricks ETL (Extract Transform Load) Pipeline

Built an end-to-end ETL pipeline in Databricks to centralize raw data, enabling scalable analytics and ML workflows.

Databricks ETL (Extract Transform Load) Pipeline

Skills

💻 Programming

Experience: 4+ years

Python R SQL C++ Shell Git MATLAB

🛠️ Data Engineering

Experience: 3+ years

PySpark Airflow Apache Spark Kafka Hive Hadoop HDFS BigQuery ETL Data Modeling Data Warehousing Feature Engineering

☁️ Cloud & DevOps

Experience: 3+ years

AWS (S3, EC2, Lambda, Redshift) Azure (Data Lake, ML Studio) Docker Kubernetes Jenkins CI/CD MLflow

🗄️ Databases & Querying

Experience: 3+ years

SQL NoSQL PostgreSQL HiveQL Google BigQuery

🧠 Machine Learning & AI

Experience: 4+ years

TensorFlow Scikit-learn Keras PyTorch FastAPI Hugging Face Transformers LangChain OpenAI API MLflow Model Context Protocol(MCP) n8n

📊 Visualization

Experience: 4+ years

Tableau Power BI Streamlit Plotly Seaborn Matplotlib

📈 Statistical Methods

Experience: 3+ years

Time Series Analysis A/B Testing Bayesian Inference Hypothesis Testing

About Me

Hi, I’m Udyan—a curious mind with a passion for building smart, data-driven systems that simplify complexity and drive meaningful change. My journey has taken me from engineering classrooms in India and Australia to Duke University in the U.S., where I recently completed my Master’s in Data Science.

Over the years, I’ve worked with teams at Accenture and Hewlett Packard Enterprise, turning complex data into real-world solutions—whether it was forecasting solar energy usage across India, developing hybrid recommendation systems for e-commerce, or building conversational AI tools to help doctors make faster decisions in intensive care units. I love solving problems that sit at the intersection of data, people, and impact.

But more than just the technical side, I thrive on collaboration. I enjoy listening, learning, and translating data into stories that others can act on. Whether mentoring new students or working across cross-functional teams, I value clear communication and thoughtful, human-centered design.

With a multicultural background and experience across India, Australia, and the U.S., I bring a global lens to collaborative problem-solving.

Udyan Sachdev