NOT YOURAVERAGEENGINEERRESEARCHERBUILDER

Software engineer turning ambitious ideas into working solutions. Currently pursuing an MS in Computer Science at Northeastern while teaching machines to learn and students to think.

Pranav PatelFormer Oracle engineer.Published researcher.Forever curious.
Scroll to begin

Index 01 · The Premise

Two years at Oracle. A paper in Springer. A 4.0 GPA. And four jobs at once.

I build at the intersection of shipping software and pushing research. After converting an Oracle internship into a full-time role on Oracle Health, and earning a Star Performer Award along the way, I went deep on the questions that wouldn't let go.

Now I split my days between graduate AI/ML coursework, a research lab, two classrooms I teach, and an advisory board. The throughline: turn ambitious ideas into things that actually work, with a researcher's rigor and an engineer's "but does it ship."

  • 4.0/4.0
    GPA @ Northeastern
  • 2
    Publications
  • 2+ yrs
    Engineer @ Oracle
  • 250+
    Students mentored

Most engineerspick a lane.I run four,at the same time.

Academia · Industry · Research · Teaching

Index 03 · The Journey

PARALLELLANES.ONE PERSON.

Most timelines stack life into a single column. Mine doesn't fit. Scroll across the lanes, Academia, Industry, Research, Teaching, Service, and watch them overlap.

20182020202220242025 · Now
Academia
MS CS · Northeastern
B.Tech IT · Manipal
Industry
Co-op · NDIF
SWE · Oracle Health · Star Performer ’23
Intern · Oracle
Research
GRA · GeoMeta + GenAI
TB Detection (CNN)
Teaching
TA · ML 6140 & DL 7150
Service
Khoury Advisory Board

The convergence

In Fall 2025, I held four roles at once.

  • Academia, MS coursework
  • Research, GeoMeta + GenAI
  • Teaching, CS 7150 & CS 6140
  • Service, Khoury Advisory Board

Key Milestones

Two papers. One in Springer. One in IJCNN.

IJCNN 2026 (WCCI, Maastricht) · Accepted Mar 2026

Simple Averaging vs. Learned Stacking in K-Fold Ensembles

Simple averaging beats learned stacking as architectural diversity grows. R² 0.89 on blind wells, +7.16 pp over the best single model.

Springer SSIC · Oct 2023

Detection of Tuberculosis from Chest X-Rays using CNN

VGG19 + custom CNN with Bayesian hyperparameter optimization. 96% on Montgomery, 86% on Shenzhen.

View publication →

Index 04 · Selected Work

BUILT TOPERFORM.

Real models, real metrics. From multi-agent medical AI to pore-pressure ensembles and medical imaging.

01 / 06 · Featured

Vita-MedAI

A 10-agent medical consultation platform. Safety screening, RAG over clinical guidelines, and reasoning, under 2 seconds end to end.

99.5%Emergency detection
  • Next.js
  • FastAPI
  • RAG
  • WebSockets
02 / 06

GeoMeta

A 5-architecture meta-ensemble for pore-pressure prediction across 11 formations, with physics-informed features and SHAP and LIME.

0.92R² on blind wells
  • PyTorch
  • XGBoost
  • SHAP
03 / 06

Brain Tumor Segmentation

U-Net segmentation of brain tumors from multi-modal MRI (T1, T1CE, T2, FLAIR), reaching clinical-grade overlap.

94%Dice coefficient
  • TensorFlow
  • U-Net
  • Medical Imaging
04 / 06

NDIF Citation Tracker

A four-phase concurrent Python pipeline unifying Semantic Scholar, OpenAlex, and GitHub to find and catalog every paper citing NDIF.

−96%Cataloging effort
  • Python
  • LLM
  • Concurrency
05 / 06

Credit Risk Prediction

Predicts loan default risk using ensemble methods with feature engineering across borrower profiles.

89%ROC-AUC
  • XGBoost
  • scikit-learn
06 / 06

Celebrity Face Recognition

A web app that detects and identifies multiple celebrity faces in an image, served through Flask.

87%Accuracy
  • OpenCV
  • Flask

MORETO COME.

A curated slice of the work. The full set spans AI/ML, full-stack, research, and hackathons.

See all work →