About Me

I'm an AI/ML engineer fluent in the language of messy data. I transform unstructured data into intelligent, scalable solutions that support automated decision-making across industries such as healthcare and finance.

With hands-on experience in LLMs, NLP, Apache Spark, Airflow, and cloud technologies like AWS, I specialize in building end-to-end pipelines and interactive dashboards that deliver real business impact.

🧠 CharanBot

Ask me anything about Sree Charan’s projects, skills, or experience!

Resume

Education

  1. University of Maryland, Baltimore County

    Jan 2024 — Dec 2025

    MS in Data Science (GPA: 3.822/4)

  2. Vellore Institute of Technology

    B.E. in Computer Science (CGPA: 8.46/10)

Experience

  1. AI/ML Engineer, OpenAI

    Jun 2025 — Present

    Designed and deployed enterprise-grade Generative AI solutions using GPT APIs, LLaMA, and Mistral, enabling automated contextual search and significantly reducing manual data discovery efforts. Built scalable RAG pipelines with LangChain, Pinecone, FAISS, and Weaviate to improve response accuracy, reduce hallucinations by 35%, and support semantic search across millions of documents. Developed high-performance AI inference services with FastAPI, Kubernetes, and model optimization techniques, reducing latency by 60% while handling over 1M daily requests. Streamlined end-to-end MLOps workflows using MLflow and AWS SageMaker, ensuring reliable production deployments and 99.9% application uptime through close collaboration with cross-functional engineering teams.

  2. Data Scientist, Nvidia

    Jun 2021 – Dec 2023

    Developed and deployed machine learning solutions using XGBoost, Random Forest, TensorFlow, and Keras to enhance predictive analytics, anomaly detection, and recommendation systems, improving model accuracy by 22% and reducing false-positive alerts by 28%. Built scalable big data pipelines with Apache Spark, PySpark, RAPIDS, AWS Glue, and S3 to process over 10TB of telemetry data daily while optimizing cloud infrastructure efficiency. Created real-time reporting dashboards in Power BI and Tableau, enabling faster business insights and stakeholder decision-making. Automated end-to-end ML workflows using Kubeflow, Jenkins, and GitHub Actions, while applying advanced feature engineering, PCA, and data optimization techniques to improve model stability and deployment reliability.

Skills

  • Python
  • SQL
  • Spark
  • Airflow
  • TensorFlow
  • PyTorch
  • Transformers
  • LLMs
  • BERT, ClinicalBERT, BART
  • Apache Kafka
  • AWS (Lambda, S3, Glue, Redshift)
  • Databricks
  • Docker
  • Flask & Django
  • Power BI
  • Plotly Dash
  • MongoDB
  • MySQL
  • LangChain
  • Java
  • C++
  • Bash
  • Machine Learning
  • Deep Learning
  • NLP
  • RAG
  • Information Retrieval
  • Time-Series
  • Scikit-learn
  • XGBoost
  • Hugging Face
  • AWS (SageMaker, EC2)
  • CI/CD
  • FAISS
  • Pandas
  • NumPy
  • Tableau
  • Matplotlib
  • Seaborn
  • FastAPI
  • PostgreSQL
  • Redis

Portfolio

AI-Powered Medical Assistant using NLP & Transformers

Built a real-time healthcare assistant using ClinicalBERT and BART. Deployed via Flask, enabling symptom triage with 72% accuracy and 1.2s latency.

Skills: NLP, ClinicalBERT, Hugging Face, Flask

GitHub |

Customer Segmentation with Apache Spark

Segmented 250K+ customers using PySpark RFM + K-Means clustering. Improved processing speed by 40% and cluster clarity by 22%.

Skills: PySpark, MLlib, Clustering

GitHub

Facial Diagnosis with Deep Transfer Learning

Web app to detect health indicators from facial images. Achieved 85% accuracy with transfer learning on medical image datasets.

Skills: CNNs, Flask, Computer Vision

GitHub

Energy Consumption Forecasting

Forecasted building energy usage using XGBoost & LightGBM. Integrated weather APIs, optimized models with GridSearchCV.

Skills: Time Series, XGBoost, Hyperparameter Tuning

GitHub

Smart Floor Cleaning Robot (Raspberry Pi)

Designed a SLAM-enabled robot using Pi + Python for real-time obstacle detection and efficient cleaning.

Skills: Raspberry Pi, SLAM, Arduino

Show Demo