Available for opportunities

Gourav Desetty

AI/ML Engineer & Software Developer — building intelligent systems with purpose. CSE student at ITER, SOA University.

scroll to explore

> Gourav Desetty
  role: "AI/ML Engineer"
  stack: [
    "PyTorch",  "LangChain",
    "FastAPI",  "Docker",
    "MLflow",   "Qdrant"
  ],
  university: "ITER, SOA",
  degree:     "B.Tech CSE",
  available:  true

01 — About

Build. Learn. Repeat.

Hello! I'm Gourav, a Computer Science and Engineering student at ITER, SOA University, Bhubaneswar. My journey in tech spans machine learning, web development, and artificial intelligence.

I have a knack for keyboards — both musical and coding. When I'm not immersed in books, you'll find me playing football or exploring new tools at the intersection of AI and software engineering.

I'm here to push boundaries, foster growth, and bring ideas to life through meaningful innovation. My journey is driven by purpose — and shared with others.

“Let's make progress possible.”

🚀

Learn Fast, Build Faster

Rapid prototyping from idea to working system with modern tools and frameworks.

🔧

Build with Intent

Every system is designed with a clear purpose — clean, maintainable, and scalable.

📈

Real-World Value

Projects that solve actual problems — from medical diagnostics to MLOps pipelines.

🧠

Stay Curious

Continuously exploring the frontier of AI, deep learning, and software engineering.

02 — Skills

My Technical Stack

From languages and ML frameworks to DevOps tools — a full-spectrum background combining specialized expertise with hands-on project experience.

Languages

PythonJavaC/C++SQLJavaScriptHTML/CSS

AI / ML / DL

PyTorchTransformersTensorFlowKerasScikit-learnOpenCV

Frameworks & Libraries

FastAPIFlaskLangChainLangGraphPandasBeautifulSoupStreamlit

DevOps & Tools

DockerGitMLflowPrometheusGrafanaDagsHub

Databases & Vector Stores

PostgreSQLMySQLPineconeFAISSQdrantChromaDB

Coursework

DSAOOPDBMSOperating SystemsMachine LearningComputer Networks

03 — Projects

Selected Work

End-to-end ML systems built with real-world constraints in mind.

Acute Lymphoblastic Leukemia CNN Pipeline

End-to-end medical diagnostic system classifying 10,661 blood smear images using fine-tuned DenseNet-121, with LangChain-powered auto-generated clinical reports.

92.5% accuracy, 0.978 ROC-AUC on imbalanced datasetWeighted loss functions to handle 7:3 class imbalanceNLP module converts inference outputs to medical reports

Model Drift Monitoring System

Reusable MLOps observability framework tracking 10K+ inferences with statistical drift detection, real-time Prometheus metrics, and a modular Streamlit dashboard.

KS-Test drift detection exposed via FastAPIGrafana dashboards with retraining alertsQdrant vector DB for embedding storage

Network Security MLOps Pipeline

Modular MLOps pipeline for phishing detection on 10K+ samples, with automated data validation, MLflow experiment tracking, and fully containerized deployment.

Best F1: 0.94 across Random Forest, GBM, AdaBoostScipy KS-Test for schema drift detectionDocker containerization for reproducible deployments