AI agent reliability · GenAI
ToolFault Atlas
An interactive reliability lab for testing how tool-using AI agents behave under malformed responses, timeouts, retries, fallbacks, and circuit-breaker conditions.
Open to Data & AI Engineering roles
I’m Deepak, a Bengaluru-based Data & AI Engineer building practical analytics, dependable data pipelines, ML reliability systems, and grounded GenAI experiences.
01 / RECRUITER LENS
Choose a role to surface the most relevant projects, strengths, and tools.
02 / SELECTED WORK
Each featured project links to working evidence—not just a description.
AI agent reliability · GenAI
An interactive reliability lab for testing how tool-using AI agents behave under malformed responses, timeouts, retries, fallbacks, and circuit-breaker conditions.
ML systems · Model monitoring
An ML reliability lab for the uncomfortable period before labels arrive—combining drift signals, conformal abstention, and champion–challenger model decisions.
Azure lakehouse · Grounded GenAI
A source-backed architecture prototype for unifying retail orders, inventory, and support data through Azure Data Factory, ADLS Gen2, Databricks, PySpark, and Delta Lake—with Power BI metrics and a grounded Azure OpenAI question layer.
Transparent scope: architecture and code prototype—not presented as a production Azure deployment.
Schedule file, API, and operational extracts with retries, parameterized dates, and auditable run boundaries.
Analytics reliability
A reproducible analysis of 146,367 flights using SQL, Excel, Power BI, Python tests, route-mix controls, and an out-of-time holdout.
Customer intelligence
K-Means customer clustering with exploratory analysis, feature preparation, notebook evidence, and documented reproducibility steps.
No projects match this view yet.
03 / CAPABILITIES
Move between insight, infrastructure, and intelligent systems without losing the business question.
From untidy inputs to decisions people can actually use.
Pipelines designed for repeatability, traceability, and scale.
Models are useful only when their behavior stays visible.
Ground the answer, test the failure mode, and keep the evidence visible.
04 / EXPERIENCE
Self-directed public builds plus hands-on experience across data quality, analysis, reporting, and machine learning.
Public, source-backed portfolio work
Building and documenting reliability-focused data and AI systems, including AfterLabel, ToolFault Atlas, Flight Reliability Lab, and an Azure lakehouse architecture prototype.
Rubixe AI
Cleaned, transformed, and validated data; performed exploratory analysis for trends and outliers; and designed interactive Power BI reports to monitor KPIs.
Lab View Academy
Prepared structured and unstructured data with Pandas and NumPy, applied supervised and unsupervised models, and evaluated K-Means clustering with the Elbow Method and Silhouette Score.
Madanapalle Institute of Technology and Science
Graduated with a CGPA of 8.53 / 10.
SELECTED CREDENTIALS
05 / ABOUT
My ECE foundation taught me to think in systems. Data work taught me to find the signal. ML and GenAI reliability taught me to question what happens after a model ships—or an agent calls a tool.
I work best where analysis, engineering, and communication meet—turning an ambiguous problem into something measurable, usable, and explainable.
06 / CONTACT
Based in Bengaluru, available immediately, and willing to relocate anywhere for the right Data & AI Engineering opportunity.