Sukumar Muthusamy · AI/ML Portfolio

Building dependable AI systems with clarity and care.

Graduate student in Artificial Intelligence (GPA 4.0). 8.5+ years in U.S. healthcare QA and automation. I design models, explain decisions (SHAP), and ship practical tools with privacy in mind.

Who I am

I’m Sukumar Muthusamy, an AI graduate student focused on reliable, interpretable models. My healthcare QA background taught me to value robustness, verification, and human-readable explanations. I enjoy turning complex pipelines into clean, auditable systems that people trust.

Where I’ve worked

Graduate Professional Assistant (GPA) — Kennesaw State University

Aug 2025 – Present · Kennesaw, GA

Building an AI assistant for Faculty Learning Outcomes alignment: maps course artifacts (syllabi, assignments, rubrics) to program outcomes using lightweight NLP and retrieval.
Designed privacy-first pipelines: on-prem data handling, PII scrubbing, and role-based access for faculty dashboards.
Delivered a prototype dashboard that surfaces coverage gaps, suggested assessments, and evidence snapshots for accreditation reviews.
Stack: Python, scikit-learn, simple embeddings/RAG, SQLite/BigQuery (pilot), Looker-style charts.

Tech Lead QA — Carelon Global Solutions (Elevance Health)

Sep 2020 – Dec 2024 · Hyderabad, India

Directed QA across 12 scrum teams; standardized workflows and ensured 100% QA sign-off before releases.
Cut production defects by ~70% via root-cause analysis and process improvements; downsized prod-support from 18→6.
Led automation with Selenium, Python, Postman; 25–30% faster regressions using Bitbucket+Bamboo pipelines.
Conducted 30+ interviews; mentored 30+ engineers; acted as domain SME across units.

Senior Software Engineer — Accenture (Client: Elevance/Anthem)

Aug 2016 – Sep 2020 · Chennai, India

Owned QA for a 10-member scrum team, from sprint planning to production deployment.
Onboarded/mentored 40+ hires; reduced transition time by ~30% with domain sessions.
Managed savings tracker reporting and high-trust internal data operations for 60+ staff.

Selected work

Insurance Claim Classifier (Explainable)

AI claim classification with SHAP; auditor-level and user-friendly rationales to support trust and compliance.

Pythonscikit-learnSHAP

Code Report

Privacy-Preserving Predictive Healthcare

Explored anonymization and differential privacy trade-offs for healthcare models; ethics-first design.

PythonDPscikit-learn

Notebook

Real-Time Cyber Threat Detection

Apache Spark Streaming + RF/XGBoost/Isolation Forest on NSL-KDD; improved detection and AUC.

PySparkXGBoostSMOTE

Code

Job Market Insights Data Warehouse

Star-schema in BigQuery; OLAP for salary/skills trends; Looker dashboards for insights.

BigQuerySQLLooker

Schema Dashboard

Iris ANN + GUI

From-scratch ANN with Tkinter UI for quick experimentation and demos.

TensorFlowTkinter

Code

Proof I keep sharpening

Certifications

CITI (AI & Human Subjects, Biomedical, Social/Behavioral), Google Cloud ML Tasks, Agile Tester, Java (ULSA), AWS Developer (former).

Awards

Multiple Impact Awards (Carelon), H&PS Zenith “Young Champion” (Accenture), Automation specialist recognitions.

Leadership & Volunteering

QQuest Secretary; UNESCO Kurukshetra organizer; Toastmasters; KSU service trip to Memphis; team-building initiatives.

Let’s connect

sukumarm1994@gmail.com · smuthusa@students.kennesaw.edu

Currently open for

AI/ML internships & research collaborations
Data projects with explainability & privacy focus
Speaking about beginner-friendly AI ethics