PORTFOLIO — VOL. 2026 BOWLING GREEN, OH

Gnana Tulasi Makineni

pronounced / gyah-nuh · too-luh-see · muh-kih-neh-nee /

I turn data into decisions teams can act on.

Data Scientist & BI Developer with 5+ years building machine learning models, dashboards, and analytics for manufacturing, insurance, and research teams.

View selected work Get in touch

Years building
data systems

10+

Production dashboards
shipped

Industries served:
manufacturing, insurance, edu

20K

NHANES records modeled
for biological-age ML

SCROLL

01 ABOUT

● Available for new opportunities

— A note from me

I build the data systems that help teams make better decisions.

I'm a data scientist focused on machine learning, business intelligence, and operational data. At Actual Reality Technologies, I build automated alert systems for manufacturing plants and scenario-modeling dashboards for insurance clients — work that has cut energy use, extended equipment life, and surfaced new savings opportunities.

Before that, I spent 18 months at BGSU Campus Operations building Power BI dashboards and data models that financial and facilities teams use every day to plan budgets and operations. My earlier research used public health data (NHANES) to train machine learning models that estimate biological age from lifestyle and medical inputs.

I care about three things: making the data trustworthy, making the answer clear, and making the dashboard fast.

Based in

Bowling Green, Ohio

Currently

Data Scientist @ Actual Reality

Education

M.S. Computer Science, BGSU

02 SELECTED WORK

Projects I've built with measurable business impact — dashboards, alert systems, and machine learning models.

Healthcare · Live Analytics · Power BI

Healthcare Claim Routing Engine

A live analytics engine for a healthcare insurance client — surfaced ~$180K in projected savings by routing claims to lower-cost in-network providers.

Developed and deployed for a healthcare insurance client. The engine evaluates claim-routing strategies against historical claim data and identifies which in-network providers would unlock the most savings. The client uses the output to back provider negotiations and make data-driven network optimization decisions.

Power BIDAXM QueryHealthcare analyticsLive data

~$180K

projected savings identified

Live

analytics engine, deployed to production

Libbey Glass · Real-time KPI · Power BI

Libbey Glass KPI Dashboard

A production KPI dashboard giving 10+ stakeholders — including the CFO and CIO — real-time visibility into Libbey Glass's manufacturing operations.

Designed, built, and shipped to production for Libbey Glass. The dashboard surfaces the manufacturing KPIs that matter most — in real time — to the people who run the business. Executive leadership (CFO, CIO) and plant operations teams all rely on it for data-driven daily decisions about the production floor.

Power BIDAXM QueryReal-time dataManufacturing KPIs

10+

stakeholders using it daily

CFO · CIO

executive leadership + plant ops

Machine Learning · Public Health Data

Biological Age Estimator

A machine learning model that estimates a person's biological age from their lifestyle and health data.

Built using NHANES — a public dataset of 20,000 health and lifestyle survey responses. I trained three different models (Random Forest, XGBoost, and ElasticNet) and compared their accuracy, then used feature importance analysis to figure out which habits and health markers matter most when the body's age drifts from the number on the calendar.

PythonRandom ForestXGBoostElasticNetPandasNHANES

20K

survey records used for training

ML models compared and tuned

Also notable —

Anomaly Detection · Manufacturing

Manufacturing Alert System

Automated alerting trained on historical signal data. Warns plant operators about unusual operating conditions before equipment fails or energy gets wasted.

↓ 10% energy · ↑ 20% equipment life

Process Optimization · Glass

Glass Production Energy Cut

Worked with the data team at a glass manufacturer to cut energy use by 4% by tuning the production process based on where the data showed waste.

↓ 4% energy use

Compliance · Power BI

Chemical Exposure Dashboards

Power BI dashboards that track chemical exposure across multiple plants. Lets the safety team compare which plants are within limits and which aren't, side by side.

Multi-plant safety view

EHS Analytics

Holiday Safety Incident Analysis

Looked into health and safety incidents around long weekends. Found a clear pattern: 30–40 more incidents happen right after holidays. The result is now used to plan safer post-holiday returns.

+30–40 incidents · pattern surfaced

BGSU · Enterprise BI

Campus Operations Dashboards

Built and maintained 5 Power BI dashboards that financial and facilities teams use every day to plan campus operations. Tuned the data models so reports load 25% faster.

↓ 25% load time · 5 dashboards

Governance · Power BI Service

Power BI Workspace Governance

Owned the operational side of Power BI at BGSU — scheduled data refreshes, managed who can see what, set up row-level security, and configured the workspace end-to-end.

End-to-end ownership

03 EXPERIENCE

2025 JAN — ONGOING

Data Scientist

Actual Reality Technologies

Developed and deployed a live analytics engine for a healthcare insurance client — identified ~$180K in projected savings by routing claims to lower-cost in-network providers.
Designed, built, and shipped a production KPI dashboard for Libbey Glass — real-time visibility into manufacturing KPIs for 10+ stakeholders including CFO, CIO, and plant operations.
Built an automated alert system for a manufacturing plant — 10% less energy use, 20% longer equipment life.
Cut energy use by 4% at a glass manufacturer by tuning the production process based on where the data showed waste.
Built chemical exposure dashboards in Power BI so the safety team can compare plant-to-plant compliance side by side.
Investigated health & safety incident data and found a 30–40 incident spike after long weekends — now used to plan safer post-holiday returns.

2023 AUG — JAN '25

Power BI Developer

BGSU Campus Operations

Built and maintained 5 Power BI dashboards that financial and facilities teams use every day to plan campus operations.
Wrote advanced DAX measures and M Query transforms to handle complex reporting needs and keep the data clean and trustworthy.
Tuned data models for speed and ease of use — 25% faster load times.
Owned the operational side of Power BI: scheduled data refreshes, managed user permissions, and set up workspace security.

2022 DEC — JUN '23

ML Research Intern

MyEdMaster

Built a machine learning model that estimates biological age from people's lifestyle and health survey answers (NHANES, 20,000 records).
Engineered features from physical, behavioral, and medical inputs and trained three different models — Random Forest, XGBoost, and ElasticNet — to compare them.
Evaluated each model with cross-validation and used feature importance to find which habits and health markers matter most for biological aging.

04 CAPABILITIES