Hello, I'm

Harsha Vardhan Reddy

AI/ML Engineer

McLean, VA (Open to Relocate)

AI/ML Engineer with 9+ years of experience designing, developing, and deploying scalable machine learning, deep learning, and Generative AI solutions. Skilled across the end-to-end ML lifecycle — data engineering, feature engineering, model building, deployment, and monitoring. Hands-on with LLMs, prompt engineering, RAG pipelines, NLP, and computer vision, building production-grade systems using FastAPI, Docker, Kubernetes, and MLOps practices on AWS, Azure, and GCP.

My Resume Contact Info

About Me

Portrait of Harsha Vardhan Reddy

AI Engineer @ Freddie Mac

I'm Harsha Vardhan Reddy, an AI/ML Engineer with 9+ years of experience designing, developing, and deploying scalable machine learning, deep learning, and Generative AI solutions. I work across the full ML lifecycle — data engineering, feature engineering, model building, deployment, and monitoring — applying supervised, unsupervised, and reinforcement learning to solve complex business problems.

I have strong expertise in deep learning architectures (CNNs, RNNs, LSTMs, Transformers) for NLP and computer vision, and hands-on experience with Generative AI — LLMs, prompt engineering, embeddings, and Retrieval-Augmented Generation (RAG). I build production-ready ML systems with FastAPI, Flask, Docker, and MLOps practices (CI/CD, model versioning, monitoring, automated retraining) on AWS, Azure, and GCP, working with large datasets using Python, SQL, Pandas, NumPy, and Spark.

9+
Years Experience
AI / Machine Learning
End-to-End
ML Lifecycle
Data → Model → Deploy
LLM + RAG
Generative AI
Production Systems
3
Cloud Platforms
AWS · Azure · GCP

Professional Experience

March 2025 – Present

AI Engineer

Freddie Mac, McLean, VA

  • Designed and developed machine learning models for classification, regression, clustering, and recommendation systems to solve real-world business problems.
  • Built and optimized ML pipelines using Scikit-learn, XGBoost, and ensemble techniques to improve model accuracy and performance, applying feature engineering, data preprocessing, and hyperparameter tuning to reduce error rates.
  • Designed and deployed end-to-end MLOps infrastructure enabling continuous integration, continuous training, and continuous deployment (CI/CT/CD) of ML models.
  • Established model monitoring and observability systems to track data drift, concept drift, and model degradation in production environments.
  • Built fault-tolerant and scalable API-based ML serving systems using FastAPI and Kubernetes, supporting high concurrent request loads.
  • Developed AI-driven recommendation systems leveraging collaborative filtering, matrix factorization, and deep learning embeddings to enhance personalization.
  • Designed and optimized NLP pipelines for large-scale text analytics, including document classification, entity extraction, and semantic search.
  • Fine-tuned transformer-based models for domain-specific applications, improving contextual understanding and response relevance.

Environment: Python, Scikit-learn, XGBoost, LightGBM, TensorFlow, PyTorch, Keras, FastAPI, Flask, Docker, Kubernetes, MLflow, Airflow, CI/CD (Jenkins/GitHub Actions).

June 2023 – Feb 2025

AI Engineer

Walgreens Boots Alliance, Deerfield, IL

  • Developed scalable chatbot frameworks capable of handling multi-turn conversations with memory and contextual awareness using LLMs.
  • Implemented distributed training strategies for deep learning models using GPU acceleration, significantly reducing training time.
  • Built automated hyperparameter tuning systems using Bayesian optimization and grid/random search to enhance model performance.
  • Designed multi-modal AI systems combining text, image, and structured data inputs for advanced decision-making applications.
  • Optimized deep learning model inference using quantization, pruning, and ONNX conversion for faster and lightweight deployment.
  • Developed intelligent alerting systems that trigger automated retraining pipelines based on performance thresholds and drift detection.
  • Built scalable batch and real-time prediction systems supporting millions of daily requests in production environments.
  • Created reusable ML microservices architecture enabling independent deployment and scaling of individual models.

Environment: Python, Hugging Face Transformers, TensorFlow, PyTorch, Keras, OpenCV, FastAPI, Flask, Docker, Kubernetes, MLflow, Airflow, ONNX, Optuna, CI/CD (Jenkins/GitHub Actions).

Oct 2021 – May 2023

ML Engineer

Merck Pharmacy, Branchburg, NJ

  • Designed and implemented scalable machine learning architectures supporting high-volume data processing and real-time inference in production environments.
  • Engineered automated ML pipelines (AutoML-style workflows) for continuous model training, evaluation, and deployment across multiple environments.
  • Built anomaly detection systems using unsupervised learning techniques to identify fraud, system failures, and unusual patterns in data.
  • Developed intelligent forecasting systems for time-series data, improving demand prediction and business planning accuracy.
  • Designed and deployed recommendation systems using collaborative filtering, matrix factorization, and deep learning embeddings.
  • Implemented model explainability techniques (SHAP, LIME) to improve transparency and stakeholder trust in ML predictions.
  • Designed A/B testing frameworks to evaluate model performance in live production environments and support data-driven decision-making.
  • Built scalable ETL pipelines to transform raw data into machine-learning-ready datasets using batch and streaming architectures.

Environment: Python, R, Scikit-learn, XGBoost, LightGBM, TensorFlow, PyTorch, Keras, FastAPI, Flask, Docker, Kubernetes, MLflow, Airflow, Optuna, SHAP, LIME, CI/CD (Jenkins/GitHub Actions).

Aug 2019 – Sept 2021

ML Engineer – Data Science

Raytheon Technologies, Waltham, MA

  • Developed end-to-end data science solutions by applying statistical analysis, machine learning algorithms, and predictive modeling techniques to solve business problems.
  • Built and deployed supervised and unsupervised learning models for classification, regression, clustering, and forecasting use cases.
  • Applied advanced algorithms including Linear/Logistic Regression, Random Forest, XGBoost, SVM, K-Means, and DBSCAN for data-driven insights.
  • Engineered distributed data processing workflows using Spark and cloud infrastructure to handle high-volume structured and unstructured datasets efficiently.
  • Built real-time analytics and prediction systems integrated with streaming data sources for instant decision-making and operational intelligence.
  • Designed advanced ensemble modeling frameworks combining multiple ML algorithms to significantly improve predictive accuracy and robustness.
  • Developed intelligent forecasting models for demand planning, sales prediction, and trend analysis using time-series and deep learning approaches.
  • Implemented explainable AI (XAI) frameworks using SHAP and LIME to provide interpretability and transparency for business stakeholders.

Environment: Python, R, Scikit-learn, XGBoost, LightGBM, TensorFlow, PyTorch, Keras, Spark, Statsmodels, FastAPI, Docker, Kubernetes, MLflow, Airflow, Optuna, SHAP, LIME.

Jan 2017 – April 2019

Data Scientist

Paradigm, Hyderabad, India

  • Designed and implemented scalable data science architectures unifying data ingestion, processing, feature engineering, modeling, and deployment into a single end-to-end analytical ecosystem.
  • Built enterprise-level predictive analytics systems capable of processing large-scale datasets to deliver real-time and batch insights for strategic decision-making.
  • Engineered distributed data processing pipelines using Spark and cloud platforms to handle high-volume, high-velocity structured and unstructured data.
  • Designed advanced ensemble learning frameworks combining multiple models to improve predictive accuracy, stability, and generalization.
  • Developed intelligent forecasting systems using statistical and deep learning techniques for demand prediction, sales forecasting, and trend analysis.
  • Built scalable feature engineering frameworks to standardize and automate feature creation across multiple data science projects.
  • Designed anomaly detection systems for fraud detection, risk assessment, and system monitoring using unsupervised learning techniques.
  • Implemented explainable AI (XAI) techniques such as SHAP and LIME to improve model transparency and stakeholder trust.

Environment: Python, R, Scikit-learn, XGBoost, LightGBM, TensorFlow, PyTorch, Keras, Spark, Statsmodels, FastAPI, Docker, Kubernetes, MLflow, Airflow, Optuna, SHAP, LIME, DVC.

Featured Projects

LLM-Powered Conversational Assistant (RAG)

Python LLMs RAG Hugging Face Transformers Embeddings FAISS / Pinecone FastAPI Docker Kubernetes

Built a scalable chatbot framework handling multi-turn conversations with memory and contextual awareness using LLMs and Retrieval-Augmented Generation. Combined vector-database semantic search (FAISS/Pinecone) with fine-tuned transformer models for domain-specific relevance, served through a fault-tolerant FastAPI + Kubernetes layer supporting high concurrent request loads.

View Code

AI-Driven Recommendation Engine

Python Collaborative Filtering Matrix Factorization Deep Learning Embeddings TensorFlow PyTorch MLflow AWS SageMaker

Developed a personalization engine leveraging collaborative filtering, matrix factorization, and deep-learning embeddings to power real-time recommendations. Backed by automated feature engineering pipelines and an end-to-end MLOps workflow (CI/CT/CD), with model monitoring for data and concept drift to maintain accuracy in production.

View Code

Anomaly Detection & Forecasting Platform

Python Scikit-learn XGBoost Time-Series Spark SHAP / LIME Airflow Kafka

Engineered unsupervised anomaly detection for fraud and predictive maintenance alongside time-series forecasting for demand planning. Built scalable batch and real-time prediction systems on Spark and Kafka streaming pipelines, with explainable-AI (SHAP, LIME) layers for transparency and automated drift-triggered retraining.

View Code

Technical Skills

Programming

  • Python
  • SQL
  • Scala
  • Java

Machine Learning

  • Regression
  • Classification
  • Clustering
  • Ensemble Models
  • XGBoost
  • Random Forest

Deep Learning

  • CNN
  • RNN
  • LSTM
  • Transformers
  • Attention Models

Generative AI

  • LLMs
  • Prompt Engineering
  • RAG Pipelines
  • Fine-tuning
  • Embeddings

NLP

  • Text Classification
  • Sentiment Analysis
  • Chatbots
  • NER
  • Topic Modeling

Computer Vision

  • Object Detection
  • Image Classification
  • OpenCV
  • YOLO

Frameworks

  • TensorFlow
  • PyTorch
  • Scikit-learn
  • Keras

MLOps

  • MLflow
  • Docker
  • Kubernetes
  • CI/CD
  • Model Monitoring

Cloud Platforms

  • AWS (S3, SageMaker, Lambda)
  • Azure ML
  • GCP AI Platform

Data Engineering

  • Pandas
  • NumPy
  • Spark
  • ETL Pipelines

Databases

  • MySQL
  • PostgreSQL
  • MongoDB
  • FAISS
  • Pinecone

Tools

  • Git
  • Jupyter
  • Airflow
  • Kafka

Achievements & Leadership

🏆

Core Committee Member

IEEE - SPS, VIT University

Served as a core committee member of IEEE Signal Processing Society, contributing to technical events and initiatives

👔

Vice President

LEO Club, VIT University

Worked as Vice President of LEO Club, leading community service initiatives and organizational activities

☁️

AWS Certified

Cloud Practitioner

AWS Certified Cloud Practitioner, demonstrating foundational knowledge of AWS cloud services and architecture

🎯

Agent Force Specialist

Certification

Certified Agent Force Specialist, showcasing expertise in specialized technical domains

Get In Touch

Let's Connect

I'm always interested in discussing new opportunities, innovative projects, and collaborative ventures in machine learning, Generative AI, LLM applications, and MLOps.

harshavardhankolla3060@gmail.com +1 (314) 433-5381 LinkedIn GitHub Portfolio