CESAR TORTOLERO Résumé (PDF)
OPEN TO AI/ML ENGINEER ROLES · NYC AREA, HYBRID OR REMOTE

Machine Learning Engineer

Cesar Tortolero
HELLO, I’M CESAR TORTOLERO

At ADP I build AI agents and LLM evaluation on AWS.

Also MCP integrations, LLM tracing, serverless APIs and MLOps.

Download résumé PDF
LAB GITHUB LINKEDIN EMAIL
Watch it get built
AGENTS, EVALS AND MLOPS AT ADP.SCROLL TO EXPLORE
01 / A LITTLE CONTEXTTHE PERSON BEHIND THE PROJECTS
C / A CLOSER LOOK

ABOUT
ME

FROM BCI RESEARCH TO PRODUCTION AGENTS.KEEP EXPLORING
AI AGENTS, LLM EVALUATION
AND MLOPS ON AWS.CESAR TORTOLERO
MACHINE LEARNING ENGINEER
ML ENGINEERING / LLM SYSTEMS / MLOPS

From EEG signals to AI agents in production.

I’m Cesar Tortolero, a Machine Learning Engineer at ADP, based in Morristown, NJ, in the New York City area. I teach developers to build, test and deploy AI agents, and debug their agent integrations when they break in production.

Before ADP, I led thoughts-to-text brain-computer interface research at Kennesaw State University. I speak English and Spanish.

CURRENTLY Pursuing an M.S. in Computer Science (Machine Learning) at Georgia Tech, part-time, while working full-time.

What I can do Experience
KENNESAW STATE · BCI RESEARCH 2021–2023ADP · ML ENGINEERING SINCE 2024
02 / WHAT I CAN DOSEVEN AREAS

WHAT I
CAN DO.

The roles and results behind each one are in Experience, below.

03 / EXPERIENCEBUILD · TEACH · SHIP · DEBUG

WORK
EXPERIENCE.

Machine learning engineering at ADP since July 2024. Before that, brain-computer interface research at Kennesaw State University.

ADP · ROSELAND, NJ

Machine Learning Engineer

– PRESENT

The 2026 work below began while I was an Associate ML Engineer and continues in this role.

2026

AI agent platform & applied AI

  • Co-delivered 18+ hands-on workshops to ~350 developers across payroll, HR, tax and analytics on building, testing and deploying AI agents. The platform now runs 250+ agents. Wrote the onboarding guide and 12 demo videos.
  • Debugged production agent integrations: MCP IAM request signing, OAuth credentials, missing deployment secrets and network rules. Brought a payroll agent and a regional integration back into service.
  • Set up LLM tracing and evaluation with Arize AX and OpenTelemetry/OpenInference. Evaluated a skill-extraction model on 2,400+ records with precision, recall, F1 and slice analysis.
  • Hardened a developer-portal API: transactional versioning, optimistic concurrency, entitlement fixes and audit identity.
  • Strands Agents
  • MCP
  • AWS Bedrock AgentCore
  • Arize AX
  • OpenTelemetry
  • OpenInference
  • NestJS
  • TypeScript
  • DynamoDB
2026

Cloud infrastructure & release engineering

  • Built a serverless cross-account network-testing API that probes across AWS accounts in parallel with bounded retries.
  • Built a network-diagnostics app that verifies firewall rules across many AWS accounts.
  • Ran 6+ weekly multi-region releases of an enterprise AI assistant platform; in one, for example, every US and EU job finished green. Wrote the runbook.
  • FastAPI
  • AWS Lambda
  • Private API Gateway
  • Terraform
  • Python
  • Flask
  • boto3
  • SSM Run Command
  • Jenkins
  • Artifactory

ADP · ROSELAND, NJ

Associate ML Engineer

–

2024 – 2025

MLOps & GenAI

  • Supported Databricks Unity Catalog and MLflow model migrations: signatures, aliases and Spark UDF batch inference, across a model inventory in the hundreds.
  • Root-caused SageMaker endpoint deployment failures in Jenkins.
  • Built GenAI playground features with Streamlit and LangChain.
  • Databricks Unity Catalog
  • MLflow
  • Spark
  • Amazon SageMaker
  • Jenkins
  • Streamlit
  • LangChain

KENNESAW STATE UNIVERSITY

Lead Researcher Thoughts-to-Text BCI

–

RESEARCH

Brain-computer interface

  • Researched decoding attempted handwriting from EEG, one character at a time, and mapped how characters separate with an autoencoder and t-SNE.
  • Presented the work at NCUR and received a $2,000 Research Excellence stipend.
  • EEG
  • Autoencoder
  • t-SNE
2022 – 2023

Research Peer Mentor

  • Mentored 50+ undergraduate researchers (2022–2023).
04 / SELECTED PROJECTSRESEARCH, EXPERIMENTS AND TOOLS

PROJECTS.

Agent tooling, EEG research and a capstone chatbot, then three smaller side projects.

All projects 06

  1. Multi-Agent Workflow01 / AGENT TOOLINGA specialist for each step. Context they can read, not change.
  2. Thoughts-to-Text BCI02 / BCI RESEARCHFrom brain signals to letters.
  3. Companionship Chatbot03 / CONVERSATIONAL AIListen for the tone, not just the words.
  4. VoiceLab04 / SIDE PROJECT · PROTOTYPEA local writing-analysis prototype that keeps text on the device.
  5. Work Helper05 / SIDE PROJECT · IN DEVELOPMENTA tool in development to help people find job opportunities and take the next step.
  6. Barking Village06 / SIDE PROJECT · FULL-STACK APPA full-stack app that puts dog-care staff workflows in one place.

Index

AGENT TOOLING

Multi-Agent Workflow

A specialist for each step. Context they can read, not change.

A multi-agent developer workflow. Specialist coding subagents split the work: one implements, one reviews, one verifies. Read-only MCP servers give them CI and source-control context, and a searchable knowledge base is compiled automatically from work sessions.

Focus
Clear roles and read-only access to delivery tools
Stack
  • MCP
  • Coding subagents
  • CI
  • Source control

No internal code, data or systems are shown. The cover is an original illustration.

Read the case study
BCI RESEARCH

Thoughts-to-Text BCI

From brain signals to letters.

Kennesaw State University research on a thoughts-to-text brain-computer interface: decoding the EEG signals of attempted handwriting into text, one character at a time. As lead researcher, I presented the work at NCUR, the National Conference on Undergraduate Research, and in two KSU symposium presentations on this work (2022), plus a related motor-imagery talk.

Role
Lead researcher, Sep 2021 – May 2023
Presented
2 KSU symposium presentations, 2022
Stack
  • EEG
  • Autoencoder
  • t-SNE

The cover is an original illustration, not research data.

Read the case study
CONVERSATIONAL AI

Companionship Chatbot

Listen for the tone, not just the words.

An LLM companionship chatbot for college students. A sentiment module names the emotion in each message so the reply can comfort or celebrate. I led the four-person capstone team and designed the backend LLM.

Award
First Place, KSU C-Day, Spring 2024
Role
Team lead, backend LLM
Stack
  • DistilGPT-2
  • Sentiment analysis
  • Python
  • React Native
  • Firebase

The cover is an original illustration. The conversation shown is sample text.

Read the case study
How I evaluate a classifier (public data)
SIDE PROJECT · PROTOTYPE

VoiceLab

A local writing-analysis prototype that keeps text on the device.

A local prototype for explainable writing-signal analysis and meaning-preserving revision. It explores patterns in language while keeping text on the local device.

Focus
Explainable analysis and thoughtful revision
Stage
Local prototype

This experiment does not establish authorship. The cover is an original illustration.

Read the case study
SIDE PROJECT · IN DEVELOPMENT

Work Helper

A tool in development to help people find job opportunities and take the next step.

A project exploring how software can help people discover job opportunities and take the next step. The current work includes a local onboarding prototype.

Focus
Making opportunity discovery more approachable
Stage
In development

The project cover is an original interface illustration, not a screenshot of a launched service.

Read the case study
SIDE PROJECT · FULL-STACK APP

Barking Village

A full-stack app that puts dog-care staff workflows in one place.

A full-stack dog-care operations project bringing staff workflows into one place: care sessions, daily planning, room status, and dog records.

Focus
Everyday operational workflows
Exploring
Clear interfaces for complex, connected tasks

The project cover is an original illustration. No customer or operational data is shown.

Read the case study
THREE FEATURED PROJECTS, THREE SIDE PROJECTS.MORE ON GITHUB
05 / LABPAGES YOU CAN USE

THE
LAB.

Three things you can try, not just read.

06 / SKILLS & EDUCATIONWHAT I WORK WITH

SKILLS &
EDUCATION.

Languages
  • Python
  • SQL
  • TypeScript
LLMs & agents
  • MCP
  • LangChain
  • RAG
  • LLM evaluation
ML & MLOps
  • scikit-learn
  • TensorFlow
  • MLflow
Cloud & infra
  • AWS Bedrock AgentCore
  • Amazon SageMaker
  • AWS Lambda
  • Terraform
  • Docker
Backend & testing
  • FastAPI
  • pytest

Education

  • Georgia Institute of TechnologyM.S. Computer Science, Machine Learning specializationONLINE, PART-TIME · EXPECTED DEC 2028
  • Kennesaw State UniversityB.S. Computer Science (Artificial Intelligence), Magna Cum Laude2021 – 2024

Spoken languages

English and Spanish FLUENT

07 / THE APPROACHHOW I THINK ABOUT BUILDING

HOW I
WORK.

Start from the person who will use it, then build.

01Start with the question.

Watch how the work gets done today before adding a feature. Checking firewall rules across our many AWS accounts took a developer about a day of manual testing, one account at a time. Asking why led to a diagnostics app that runs the same checks from one screen in minutes.

  • Watch the work today
  • Ask why
  • One screen
See the experience
02Build something tangible.

A working version makes an idea easier to question. For my capstone, the companionship chatbot ran as a real app with logins, chat and past conversations, and its sentiment module was tested by feeding it labelled prompts and checking the emotion it named.

  • A real app
  • Labelled prompts
  • The emotion named
See the work
03Stay curious. Keep refining.

The details are part of the work. Each of the 18+ workshops showed where developers got stuck, such as missing secrets, access groups and network rules, and the onboarding guide changed to match. It has been revised more than 40 times.

  • Missing secrets
  • Access groups
  • Network rules
Say hello

Still learning.
Still making.