Hitesh Reddy Buereddy

Who am I?

I'm Hitesh Reddy, a student passionate about tech and problem-solving. I began with C++, mastering OOP and solving 200+ LeetCode problems, which sharpened my skills in DSA. I expanded into Python, SQL/NoSQL, and tools like Pandas, NumPy, Tableau, and Matplotlib to turn data into insights.

I explored web development with HTML, CSS, JavaScript, React, and Django, building full-stack apps and using Git/GitHub for collaboration. Currently, I'm interning in the field of Generative AI, where I work on large language models, prompt engineering, and building intelligent systems that generate human-like content and solve real-world tasks.

Alongside this, I'm deepening my expertise in machine learning using TensorFlow, Keras, Scikit-learn, and PyTorch. I'm driven by continuous learning and a passion for building impactful, AI-powered solutions as I work toward a software or data-focused role.

Timeline

July 2025 - Present

AI/ML Engineer

Paychex

Built PySpark and Airflow pipelines on AWS (S3, RDS, Lambda, Kafka) to process 12TB+ payroll/benefits data weekly, delivering compliance-ready reports faster. Trained and deployed ML models (TensorFlow, PyTorch, Scikit-learn) to forecast attrition and detect anomalies, preventing 9,000+ payroll discrepancies annually. Orchestrated MLOps with MLflow, Docker, Kubernetes, and CI/CD (Jenkins, GitHub Actions), integrating models into production and visualizing 150+ KPIs in Tableau/Power BI for audits.

Sept 2023 - May 2025

Master's Degree

New Jersey Institute of Technology

I moved from India to the U.S. to pursue my Computer Science degree at NJIT, driven by a passion for technology and innovation. With a 4.0 GPA, I've honed my skills through projects like TimeFlex, Image Classification, and coursework in Machine Learning, Big Data, and Operating Systems.

2021 - 2023

AI/ML engineer

Airbnb India

Engineered large-scale ETL pipelines with PySpark, Airflow, and Kafka to process 50M+ records daily, reducing reporting latency by 18 staff hours per week. Built and deployed predictive and deep learning models (TensorFlow, PyTorch, Scikit-learn) for occupancy forecasting, demand prediction, and anomaly detection, generating $25M+ in booking value and improving revenue projections. Applied NLP (spaCy, NLTK, Hugging Face) on 3M+ reviews to uncover service issues and boost 15K+ positive ratings annually. Deployed ML solutions on AWS (Lambda, S3, RDS) with MLflow, optimized PostgreSQL/MongoDB for 10TB+ data, and led MLOps workflows with Docker, Kubernetes, and CI/CD, while mentoring junior analysts on scalable deployments.

Jan 2020 - Feb 2023

Working Body Member

The Robotics Club

I volunteered for 3 years in the Robotics Club at SNIST, serving as a Technical Coordinator and eventually the Technical Head for the Roboved fest. I contributed to organizing the event, raising funds, and guiding 7 members in building their first robot. Our efforts led to a successful event that fostered technical learning and hands-on experience for participants.

Aug 2019 - June 2023

Bachelor's Degree

Sreenidhi Institute of Science and Technology

I completed my Bachelor's degree in Computer Science from Sreenidhi Institute of Science and Technology with a CGPA of 8.73/10. In addition, I served as a class representative for 2.5 years, collaborating with professors to manage academic activities and address student concerns.

Featured Projects

AI Driven Negotiator for Customer Lending

Developed and deployed an LLM-powered AI assistant using LLaMA 3.3, with custom function calling and multi-agent workflows via LangGraph to enable dynamic, context-aware conversations.
Performed advanced prompt engineering to refine model behavior across diverse inputs, reducing hallucination and improving response consistency and relevance.
Designed and executed structured model evaluations using DeepEval, measuring precision, recall, hallucination rate, and bias; automated test cases integrated into CI pipelines.
Conducted functional, performance, and ethical testing (latency, fairness, toxicity), validating inference outcomes, preprocessing logic, and end-to-end reproducibility
Collaborated with data scientists to define AI feature acceptance criteria, and containerized services with FastAPI, PostgreSQL, and Docker Compose for scalable deployment.

Plan Genie

Built Plan Genie, a full-stack productivity app using React.js, Django, and PostgreSQL, enabling users to securely log in and manage personalized TODO lists.
Integrated LLaMA 3.3 to generate AI-powered learning plans based on user-entered topics, with an approval workflow to curate and customize suggested tasks.
Implemented secure user authentication (login/signup) and dynamic task management with real-time updates, providing a personalized and interactive user experience.

TimeFlex Trader

Developed hybrid deep learning models (LSTM & Transformer) for stock price prediction with 96% R² and minimal RMSE
Enhanced predictions using rolling statistics and volatility metrics.
Built a data pipeline for preprocessing and normalization to optimize forecasting performance.

VisionNet

Developed and compared a custom CNN with ResNet-18 on the CIFAR-10 dataset for image classification.
Achieved 80% accuracy with custom CNN and 95% with ResNet-18, demonstrating the efficacy of deeper architectures.
Employed data augmentation and optimization techniques to enhance model performance and generalization.

AgriProtech

Constructed a custom made Convolutional Neural Network (CNN) model for predicting plant diseases based on leaf characteristics which achieved 98% accuracy
Integrated the model into a web application using Flask and HTML, providing users with disease predictions and suggestions for appropriate medicines and pesticides.

Cardiopathie

Designed CNN models (Inception ResNet v2, VGG, ResNet-50) achieving 94% segmentation accuracy on 2400 CT images, classifying calcium-positive and calcium-negative cases using NRRD medical datasets.
Developed a pipeline validated on 20,084 CT scans, achieving >90% precision in calcium score prediction, significantly improving diagnostic speed and reducing manual intervention.
Leveraged TensorFlow and SimpleITK for efficient preprocessing and segmentation, reducing computation time by 30% on 4GB GPU hardware, ensuring scalable clinical deployment.

SkyWeathercast

Component Architecture: The app uses React components like WeatherCard and SearchBar, managing state with useState and triggering API calls via useEffect.
API & Error Handling: Weather data is fetched from weatherapi.com using fetch or axios, with try/catch for error handling.
Geolocation & Storage: The app uses Geolocation API for location-based weather and localStorage to persist user data.

Cocinero

Autonomous Cooking: COCINERO automates food preparation, chopping ingredients, adding seasonings, and mixing them based on user inputs.
Convenience: Designed for busy individuals, it simplifies cooking for working professionals, students, and bachelors.
Focus on Salads: Initially, it automates salad preparation, including cutting, seasoning, and mixing for a tasty dish.

Cognute AI

Cognute Ai

Working on it
Will update once done

Data Analysis using tableau

working on it
Will update once done

ToDo and Chatapp

Todo and chatapp

working on it
Will update once done

Portfolio chatbot

working on it
Will update once done

Skills & Expertise

Python

C++

JavaScript

LLM's

Prompting

AWS

Jenkins

Git/Github

Docker

Django

React

Mysql & PostgreSQL

HTML5

CSS3

TensorFlow

PyTorch

Scikit-Learn

Data Analysis

Let's Connect

I'm always ready to Collaborate and learn from others. Feel free to reach out to me via email or LinkedIn.

buereddyhiteshreddy28@gmail.com

+1 (929)-498-6356

H.A.R.I.S

Hi, I'm H.A.R.I.S, (Hitesh's AI Resource and Information System). What are you searching about hitesh?