Skip to content
View adullagayathri's full-sized avatar
1 follower · 1 following

Achievements

Achievement: Pull Sharkx2Achievement: YOLOAchievement: Quickdraw

Achievements

Achievement: Pull Sharkx2Achievement: YOLOAchievement: Quickdraw

Highlights

  • Pro

Block or report adullagayathri

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
adullagayathri/README.md

Hi, I'm Gayathri Adulla 👋

AI Engineer building production Computer Vision, Multimodal AI, and Machine Learning systems.

🔭 Currently building computer vision pipelines for automated housing-code-violation detection at Third Estate Analytics, processing GPS/video telemetry and VLM-based classification at 90% accuracy across a 300K-parcel dataset.

🎓 M.S. in Artificial Intelligence — University at Buffalo (2025–2026)

🌱 Currently exploring: Retrieval-Augmented Generation (RAG), LLM applications, and agentic AI systems

📫 Reach me: LinkedIn · Portfolio · gayathri.adulla.18@gmail.com

🛠️ Tech Stack

Languages: Python · SQL · Java ML / DL: PyTorch · TensorFlow · Scikit-learn · Transformers (DistilBERT, VLMs) Computer Vision: OpenCV · ResNet · SMPL-X · Active Learning pipelines Infra / MLOps: Docker · Kubernetes · FastAPI · Redis · AWS · GCP Data: PostgreSQL · MongoDB · Spark · Hadoop · ETL pipelines

📌 Featured Projects

ProjectDescriptionStack
Multimodal Movie Genre PredictionFused text (DistilBERT) + image (ResNet-18) deep learning model for multi-label genre classification. Live demoPyTorch, Transformers, ResNet
Deepfake Detection (β-VAE)Hybrid generative + discriminative deepfake detection — 98% accuracy, 0.997 ROC-AUC, on FFHQ + Stable Diffusion dataPyTorch, VAEs, Scikit-learn
3D Human Mesh ReconstructionSMPL-X-based 3D human reconstruction & rendering pipeline from monocular videoPyTorch, Trimesh, Pyrender, Gradio
AI Agent Microservice PlatformDistributed orchestration runtime simulating 50+ concurrent AI agents with health checks & retry logicFastAPI, Docker, Redis

📊 GitHub Stats

Gayathri's GitHub statsTop Langs

Pinned Loading

  1. multimodal-movie-genre-predictionmultimodal-movie-genre-predictionPublic

    Multimodal deep learning model predicting movie genres by fusing DistilBERT text embeddings with ResNet-18 visual features from posters. Includes live web demo.

    Jupyter Notebook 1

  2. 3D-human-mesh-rendering3D-human-mesh-renderingPublic

    3D human body mesh reconstruction from monocular video using SMPL-X motion parameters, with a CPU-optimized rendering pipeline and Gradio-based comparison UI. Live demo available.

    Jupyter Notebook 1

  3. Deepfake_DetectionDeepfake_DetectionPublic

    Hybrid β-VAE + supervised classifier for deepfake face detection — combines generative latent-space modeling with discriminative learning. 98% accuracy, 0.997 ROC-AUC.

    Jupyter Notebook

  4. Role-Based_E-Commerce_Management_SystemRole-Based_E-Commerce_Management_SystemPublic

    Java desktop application managing multi-role e-commerce workflows (Admin, Warehouse, Delivery, Fraud Analyst) with role-based dashboards, Swing UI, and MySQL backend.

    Java

  5. AI-Agent-Microservice-PlatformAI-Agent-Microservice-PlatformPublic

    Containerized microservice runtime orchestrating 50+ concurrent autonomous AI agents with health monitoring, retry logic, and structured logging via FastAPI and Redis.

    Python 1

  6. Housing_code_violationsHousing_code_violationsPublic

    Production computer vision pipeline converting GoPro dashcam footage into address-resolved housing code violation reports — GPMF telemetry processing, SSIM keyframe extraction, and Gemini Vision VL…

    Python