Recruiter Packet
Role fit, stack, resume, and submission-safe proof links in one place.
View recruiter packet →Choose the proof path
Hiring Manager Proof
Architecture, validation, failure modes, tests, and evidence-backed answers.
View manager proof →Use-Case Labs
Small governed data workflows that turn messy inputs into reviewable outputs.
View use-case labs →Why this is different
This portfolio is built like an evidence system, not a brochure. Public claims route through a versioned evidence ledger, generated manifest, policy checks, and unsupported-claim rejection.
That is the same engineering pattern shown in the labs: validate inputs, preserve provenance, route exceptions, expose boundaries, and keep automation reviewable.
Experience snapshot
Lead Software Engineer (W2 Contract) — Universal Creative / NBCUniversal (most recent publicly listed role)
- Governed ingestion: Python document-ingestion harness and checkpointed Microsoft Graph/SharePoint delta-sync with parser provenance and review bundles.
- ACCEPT/QUARANTINE/SYSTEM-failure routing and guarded ClickUp automation, preserving owner approval over every status change.
Software Engineer II (Data Engineer) — 4C Strategies North America
- Idempotent Kafka-to-PostgreSQL consolidation service with replay handling, DTO/schema validation, and paginated APIs.
- Grafana/Prometheus/Loki observability and GitLab CI contract-test gates (JUnit 5, Mockito, Spring Cloud Stream test binder).
Earlier roles: Senior Cloud Solutions Architect at Accenture (Microsoft Advanced Azure Engagement) and Senior Data Analyst / Data Engineer at Flexport.
Recruiter Snapshot
- Target roles: Senior Data Platform Engineer, AI Ingestion Engineer, Streaming Data Engineer
- Work model: Remote, Orlando hybrid, Central Florida hybrid
- Core stack: Azure, Kafka, Python, SQL, PostgreSQL, Microsoft Graph, SharePoint, CI/CD
- Proof: evidence-backed chat, GitLab repos, resume
Hiring Manager Proof
- Architecture
- Tests
- Evidence
- Operations
Use-Case Labs
Small data engineering applications that show how messy signals become governed decisions. Synthetic or public data only.
Patient Signal Intake Lab
Synthetic care-event intake: schema validation, quarantine routing, review packets.
View lab →Grid Event Quality Monitor
Synthetic grid/meter telemetry: timestamp validation, drift detection, degraded-data handling.
View lab →Field Change Order Evidence Tracker
Synthetic field change requests: document intake, duplicate detection, owner-approval boundary.
View lab →Evidence-Grounded RAG Evaluator
How this site's own chat rejects unsupported claims before you ever see them.
View lab →Engineering Principles
The same principles that govern the labs and this site's own chat assistant.
- Determinism
- Provenance
- Observability
- Governance
Public Reference Projects
Personal, clean-room reference implementations — synthetic data only, no confidential employer code.
Acceptance Harness
Deterministic, infrastructure-free Java 17 validation harness for XMPP-sourced message ingestion.
View repo →Air-Gap Messaging Ingest & Observability
Idempotent Kafka-to-PostgreSQL ingest pipeline with Grafana/Prometheus/Loki observability.
View repo →ShopLabs Lakehouse Lab
Clean-room D2C commerce lakehouse using Python and DuckDB with synthetic, seedable order data. No real customer or payment data.
View repo →CareerAssistAI Scoring Pipeline
Clean-room FastAPI service scoring resume/job-description overlap with a transparent rationale. No private tracker data.
View repo →Certifications
- Azure Data Fundamentals 11/2025
- Azure AI Fundamentals 10/2025
- Databricks Platform Architect 11/2025–11/2026
- Confluent Data Streaming Engineer 09/2025–09/2027
Connected Public Signals
Public, no-login signal adapters: bounded inputs, caching, privacy limits, safe degradation.
How Answers Are Generated
Evidence-backed answers
Ask one of these — every answer includes sources and a metadata strip (evidence gate, safety gate, confidence).
Unsupported-claim example — “Did Nkosi lead a Fortune 100 production AI rollout?”
Safe rejection: no grounded public evidence supports that claim; the assistant answers about documented governed ingestion, Graph/SharePoint, Kafka/PostgreSQL, and portfolio labs instead.
Evidence gate: rejected · reason: unsupported public evidence