Open to Data Reliability / ML Data Engineer roles · Lynnfield, MA (Boston area)

I keep bad data out of ML models.

I'm Aria Ushani, a data-quality engineer with ten years inside production pipelines at Melissa Data: SQL validation at scale, ETL fixes at the source, and automated data checks in Python and C#. M.S. in Computer Science with a deep-learning thesis.

Download résumé Email me

10+

years in production data

3.87

GPA, M.S. Computer Science

17K

images in my ResNet50 thesis dataset

languages: Python, SQL, C#, Java

Where I fit

Most model problems are data problems. I work at the gate between raw data and training.

i. IngestAPIs & data servicesTested and load-tested with Postman and JMeter
ii. ValidateSQL integrity checks at scaleMy core strength, across large production pipelines
iii. TransformETL fixes at the sourceSSIS on SQL Server, with engineering
iv. ShipCI/CD quality gatesBuild validation; Docker and Kubernetes
v. TrainDeep learningResNet50, transfer learning, augmentation

Case studies

Melissa Data · QA Engineer, data pipelines · 2015 — now

Data integrity at production scale

Designed SQL validation strategies to verify data integrity across large production pipelines.

Result: integrity verified across large-scale production pipelines

Fixing data at the source

Analysed data-quality issues with engineering, implemented fixes in ETL processes, and documented the pipeline logic.

Result: fixes made in the ETL itself, documented in JIRA and Confluence

New data processing, proven first

Led proof-of-concept projects evaluating new data processing approaches.

Result: demonstrated performance improvements and business impact

Automated checks for data services

Built Python and C# frameworks for REST APIs, ETL and data services; Postman and JMeter suites for high-volume load; CI/CD build validation.

Result: reliability under high-volume workloads, with failures tracked in JIRA

Machine learning

Hands-on with the model side, so I know what training data has to look like.

M.S. thesis · UMass Dartmouth · Dec 2024

Blood cell classification with ResNet50

Transfer learning and data augmentation on BloodMNIST: 17,092 images across 8 blood-cell types.

M.S. Computer Science

University of Massachusetts Dartmouth, GPA 3.87.

Data: SQL · SQL Server
SSIS (ETL)
XML / JSON

Quality: Data validation
Integration · regression
Load (JMeter)

Code: Python · C#
Java

Platform: Docker · Kubernetes
Git · CI/CD
JIRA · Confluence

ML: CNNs · ResNet50
Transfer learning
Augmentation

2015 — nowQA Engineer, Melissa DataData pipelines, ETL, SQL validation, test automation

2024M.S. Computer ScienceUMass Dartmouth · GPA 3.87

2015IT Technician, CarboniteServers, storage, networks