Venkatesha Nayak.

Open to full-time roles · Bengaluru, India

From clinical practice to AI engineering.

I’m Venkatesha — an Ayurvedic doctor moving into AI engineering. I like building machine learning systems, data pipelines and healthcare applications that solve real problems.

EmailGitHubLinkedIn
View projectsRésuméGet in touch
01 — About

Background.

Dr. Venkatesha Nayak B
Bengaluru, IndiaBAMS · PGDM

I’m a doctor turned AI engineer, combining clinical knowledge with data science to solve healthcare problems.

After practicing as an Ayurvedic physician, I transitioned into AI & Data Science through my PGDM at IIHMR Bengaluru. During my internship at the Data, Evidence & Action Lab, I worked on machine learning models, healthcare APIs, and analytics platforms for public health research.

I enjoy building products where medicine and technology meet, especially when they can simplify complex healthcare problems using data.

Physician (BAMS)PGDM · IIHMR-BPythonMachine LearningMLOpsHealthcare AI
02 — Experience

Roles.

  1. Apr 2026 – Jul 2026Bengaluru

    AI Engineer Intern

    Data, Evidence & Action Lab

    • I built an analytics platform that processes India’s national health survey data automatically and publishes reproducible results — no one has to re-run anything by hand, and reruns can’t quietly change the numbers.
    • I built Precision MPI, the lab’s flagship model, which estimates poverty for ~589,000 villages — three administrative levels below where the survey actually collected data.
    • I recomputed India’s national poverty index from scratch and checked it against the government’s official numbers. It came out within a rounding point of theirs, which is what made the village-level work credible.
    • I trained and deployed child undernutrition risk models for two age groups, and made them explainable — a prediction comes back with the reasons behind it, not just a score.
    • I built the lab’s public analytics site — interactive charts and district maps that anyone can explore, updating themselves from the data pipeline.
    • I added an LLM layer that drafts plain-language policy summaries for each map and dashboard, cached so the same view never costs twice.
    PythonMachine LearningGeoPandasFastAPIReactAWSDockerSQL

Before that — clinical practice

Feb 2025 – Jul 2025

Ayurvedic Physician

Sapthashwa Ayurvedic Health Care Center

OPD consultations and Panchakarma therapies for 6–7 patients a day — treatment plans, follow-ups and the records that go with them. Where I learned what clinical data looks like before anyone cleans it.

Oct 2023 – Oct 2024

Ayurvedic Intern

SDM Institute of Ayurveda & Hospital

A year of OPD and IPD clinical training — patient assessment, diagnosis, patient education and hospital records.

03 — Selected Work

Selected projects.

Six systems across healthcare AI, data engineering and machine learning. Open any case study for the problem, the approach, the outcome, and what went wrong on the way.

Precision MPI — Village-level MPI — Shivamogga district, 1,444 villages
Village-level MPI — Shivamogga district, 1,444 villages
01DEA LabFlagshipLive

Precision MPI

Poverty, estimated for every village in India.

Estimates for ~589,000 villages that hold up on held-out survey clusters and track the national rural figure when rolled back up.

Python · scikit-learn · GeoPandas · AWS S3 · Leaflet

Data, Evidence & Action Lab
Healthcare Analytics Dashboard — Study page — child malnutrition in Karnataka
Study page — child malnutrition in Karnataka
02DEA LabFrontendLive

Healthcare Analytics Dashboard

NFHS-5 findings, mapped and explained.

The lab’s public front door — national survey findings anyone can explore without asking for the data.

React (Vite) · Recharts · Leaflet · GitHub Actions

datatoactionlab.org
TrialScope — Therapeutic-area dashboard — obesity & metabolic disorders
Therapeutic-area dashboard — obesity & metabolic disorders
03Full-stackLive

TrialScope

Competitive intelligence for clinical trials.

~13,000 trials tracked and refreshed daily; a whole new disease area needs config and data only.

Python · FastAPI · PostgreSQL · Next.js · TypeScript

Live demoGitHub
Prakriti AI — Assessment entry — Vata / Pitta / Kapha
Assessment entry — Vata / Pitta / Kapha
04Clinical systemsLive

Prakriti AI

Two questionnaires, and the gap between them.

One focused, seasonally-aware plan instead of a wall of generic advice.

React 19 · Vite · Tailwind CSS · Recharts

Live demo
Hospital Admissions Forecasting — Forecast dashboard — 7-day admissions horizon
Forecast dashboard — 7-day admissions horizon
05Time seriesLive

Hospital Admissions Forecasting

Forecasting demand from sparse, messy data.

Shipped the simpler model — it degraded gracefully on quiet days where the seasonal one swung wildly.

Python · Pandas · Time-series forecasting

Live demoGitHub
  1. Versioned data
  2. Train
  3. Containerise
  4. Deploy
  5. Monitor
  6. Retrain

The loop that keeps it running

06MLOpsOpen source

Logistics Delay Prediction

Shipment delays, predicted and monitored.

A pipeline that retrains, redeploys and reports on itself without me in the loop.

LightGBM · FastAPI · Docker · DVC · Grafana

GitHub
04 — Capabilities

What I work with.

Grouped the way I actually use them — the language, the models, the data underneath, and everything needed to get a model in front of someone.

Languages

  • Python
  • SQL
  • JavaScript
  • TypeScript
  • R

Machine Learning

  • scikit-learn
  • XGBoost
  • LightGBM
  • TensorFlow
  • PyTorch
  • SHAP
  • Classification
  • Regression
  • Time-series forecasting
  • Semi-supervised learning

Generative AI & LLMs

  • LLM applications
  • Llama-3 · Groq
  • Prompt engineering
  • RAG
  • Inference cost optimisation

Data & Geospatial

  • Pandas
  • GeoPandas
  • PyArrow / Parquet
  • PostgreSQL
  • MongoDB
  • AWS S3 (boto3)
  • ETL pipelines

MLOps & Deployment

  • FastAPI
  • Docker
  • Git
  • GitHub Actions
  • MLflow
  • DVC
  • Weights & Biases
  • Prometheus
  • Grafana
  • Render
  • Vercel

Frontend & Visualisation

  • React
  • Vite
  • Next.js
  • Recharts
  • Leaflet
  • Streamlit
  • Power BI
  • Tableau

Healthcare Domain

  • Clinical assessment
  • Diagnosis & treatment planning
  • EMR data
  • NFHS / DHS survey data
  • Multidimensional Poverty Index
05 — Currently

What I’m learning next.

Most of what I know came from building things I hadn’t built before. This is the current list.

Exploring

  • LLMs for healthcareWhere language models are actually safe to use on clinical text — and where they are not.
  • MCP serversGiving models real tools instead of prompt-stuffed context.
  • ML system designThe parts that break in production rather than in training.
  • Healthcare AI agentsMulti-step workflows over health data, with a human in the loop.
  • Clinical NLPExtracting structure from notes written for humans, not machines.

Recent repositories

All repositories on GitHub
06 — Education & Recognition

Credentials.

Education

PGDM — AI & Data Science in Healthcare

Indian Institute of Health Management Research, Bengaluru (IIHMR-B)

Jul 2025 – Present · Expected 2027

Bachelor of Ayurvedic Medicine & Surgery (BAMS)

SDM Institute of Ayurveda and Hospital, Bengaluru

Oct 2018 – Oct 2024

Certifications

Project Management for Managers — NPTEL SWAYAM

12-week Elite certification · Top 5% rank

NPTEL

Six Sigma — NPTEL SWAYAM

DMAIC process improvement & quality optimisation

NPTEL

Data Literacy Fundamentals — DataCamp

Data interpretation, analytics & data-driven decisions

DataCamp

Languages

Konkani, Kannada — NativeEnglish — Full ProfessionalHindi — ConversationalTamil, Telugu, Malayalam — Elementary

07 — Contact

If you’re building in healthcare, let’s talk.

Looking for opportunities in AI engineering, healthcare data science and machine learning.
Bengaluru, India · available immediately.

Email meDownload Résumé
venkateshnayakb23@gmail.comgithub.com/venkateshanayakbLinkedIn+91 94805 60737