08/19/2026 updated

JA
100 % available

Senior Software Engineer | Backend Systems, Applied AI and Reliability

USA
Only remote
B.S. Computer Science and Engineering, MIT | 2014-2018
USA
Only remote
B.S. Computer Science and Engineering, MIT | 2014-2018

Profile attachments

jon-atkins-cv.pdf

About me

Senior software engineer with about eight years across enterprise software, payments, logistics and applied AI. I own backend systems from design through deployment and on-call support, with particular depth in database changes, failure handling and production diagnostics.

OpenTelemetryJava (Programming Language)JavaScriptTest AutomationPython (Programming Language)PostgreSQLSoftware ArchitectureRedisPrometheusTypeScriptGrafanaApache KafkaDocker
Backend systems, applied AI and reliability

I build and operate backend systems, especially where database changes, failure handling and production diagnostics matter. Recent work includes payments, logistics, enterprise software and AI-assisted operations.

Languages
TypeScript, JavaScript, Java, Python and SQL.

Data and AI
PostgreSQL, Redis, Kafka, OCR, document AI, structured model outputs and evaluation harnesses.

Cloud and delivery
Docker, GitHub Actions, container release gates and repeatable builds.

Production engineering
SLI and SLO design, OpenTelemetry, Prometheus, Grafana, incident response, database migrations, rollback planning and runbooks.

Typical work includes idempotent workflows, transactional outbox patterns, large backfills, partitioning, expand-contract migrations, cursor-based APIs, confidence thresholds and human review for AI workflows.

Experience

Professional experience5 - 10 years
Freelancer since2026

Languages

EnglishNative speaker

Project history

Independent Software Engineer

Independent practice
Jan 2026-present. Deliver fixed-scope backend audits, migrations and implementations while keeping existing systems usable. Recent work includes a 75-million-row PostgreSQL partitioning migration with dual writes and checksum verification; an evaluation harness for AI-assisted operations with frozen test cases, citations, schema validation, budgets and human review; and a TypeScript workflow service with idempotent jobs, bounded retries, a dead-letter queue, Docker packaging and release gates.

Senior Software Engineer

Flexport
Apr 2024-Dec 2025. Owned the backend integration between freight-document ingestion and account workflows. Led AI-assisted ingestion using OCR, schema-constrained outputs, confidence thresholds and human review. Replaced noisy alerts with latency and error-budget alerts, led an expand-contract migration from a shared PostgreSQL schema to service-owned data, and mentored two engineers into service ownership.

Senior Software Engineer

Stripe
Oct 2021-Mar 2024. Technical lead for five backend engineers covering merchant onboarding, settlement reporting and internal payment operations. Added idempotent consumers and a transactional outbox, ran a 120-million-row ledger backfill behind dual reads and checksums, coordinated an incident rollback for connection-pool exhaustion, and cut p95 search latency from 820 ms to 240 ms with cursor-based queries and targeted indexes.

Software Engineer II

Oracle
Jul 2020-Sep 2021. Owned the reconciliation service for an enterprise billing product using Linux, Java, Python and Oracle Database. Reworked a full-table nightly job into checkpointed incremental batches, cutting processing for about 12 million rows from over three hours to 47 minutes. Designed a backward-compatible schema change and became the release-week escalation contact for reconciliation defects.

Software Engineer

Oracle
Jul 2018-Jun 2020. Built billing configuration and support tooling on an eight-person product team. Added JUnit regression fixtures from production defects, created a Java diagnostic command for configuration, database and JVM state, and documented recovery steps for recurring batch failures.

Software Engineering Intern

Oracle
Jun-Aug 2017. Built a Java and SQL validation utility for nightly reference-data imports. Row-level failure reports reduced a typical support investigation from several hours to about 20 minutes. Added regression fixtures and setup notes so the tool could be maintained after the internship.

Contact form

Log in to get in touch

You need to be logged in to use the contact form.

Sign upLog in