My Profile

My Profile

👋 Hi, I’m Kalyan Narayana - WORK IN PROGRESS

Aspiring Data Scientist & Full-Stack Engineer focused on ML systems, product analytics, and production-grade pipelines.
I design, build, and ship: from EDA & modeling → APIs & dashboards → MLOps & monitoring.

Current interests: GenAI apps, Prompt engineering, Scalable DSML pipelines, and Aesthetic UI/UX experience (in progress).

🚀 Snapshot

  • 🎯 Core: Python, SQL, ML, DS workflows · FastAPI · React/Streamlit · Docker · CI/CD
  • 🧠 ML: Supervised/Unsupervised · Time Series · NLP · Recommenders · Model Monitoring
  • 🏗️ MLOps: DVC · MLflow · Kubeflow/GKE · Seldon/Istio · Evidently · GitHub Actions · Terraform
  • 📊 Product Analytics: Experimentation (A/B & Bandits), retention/funnel, North-Star metrics
  • 🌐 Cloud: GCP, AWS, plus edge/Raspberry Pi for quantized deploys

🧩 What I’m Good At

Languages: Python · SQL · TypeScript/JS · Bash
Libraries: pandas, NumPy, scikit-learn, XGBoost, PyTorch, StatsModels, SHAP
Data & Viz: Matplotlib, Plotly, Altair, Seaborn (when allowed), Grafana
Web & APIs: FastAPI, Flask, React, Next.js, Tailwind, Streamlit, Dash
Data Ops: Postgres, BigQuery, Airflow, dbt, Feast (Feature Store)
Infra: Docker, Kubernetes, Terraform, GitHub Actions, Prometheus

Value I bring: clear problem framing, reproducible analytics, and pragmatic engineering.

🛠️ Selected Work

  • Real-Time Energy Forecasting — LSTM + DVC + MLflow + Streamlit, with monitoring & alerts
    Repo: 🔁 [link] · Demo: 🔁 [link]
  • Retail Demand Forecasting (XGBoost) — multivariate features, SHAP insights, deployment to GCP
    Case study: 🔁 [link]
  • GenAI Product Assist — RAG + evaluation harness; prompt strategies (zero/few-shot, CoT)
    Demo: 🔁 [link]
  • A/B Testing Toolkit — classical & Bayesian, sequential monitoring, power analysis
    Docs: 🔁 [link]

📚 Teaching & Writing

  • Curricula created across ML, Stats, SQL, Python, Prod Analytics, MLOps
  • Guides on bias–variance, GMM/DBSCAN/HAC, time series (SARIMA/Prophet), LangChain
  • I enjoy turning messy topics into clean, reproducible notebooks and actionable playbooks
    Read: 🔁 Blog /blog · 🔁 Notebooks /projects

🗺️ Speaking & Community

  • Talks: From Notebook to Prod: A DSML Playbook, Prompt Engineering That Ships
  • Mentoring: beginner → advanced DS/ML roadmaps; portfolio reviews; capstone scaffolding

🧱 Principles I Work By

  • Clarity first. Define the metric, then the model.
  • Automate the boring. Reuseable pipelines > ad-hoc scripts.
  • Ship small, learn fast. Instrument everything, close the feedback loop.

📨 Contact

🎖️ Quick Badges

Python scikit--learn PyTorch FastAPI PostgreSQL Docker Kubernetes GCP AWS

🧭 Site Notes (Theme Tips)

  • Chirpy: put this file as /about.md. Headshot at /assets/img/profile/kalyan.jpg.
  • Minimal Mistakes: use layout: single or page, and place under /_pages/about.md with permalink: /about/.
  • Add social links in _config.yml (Chirpy: social:; MM: author: map).