Selected work

An enterprise security product shipped a syslog writer/parser/forwarder with serious liabilities: an unsupported, end-of-life Python runtime, an ancient base container (Debian 7), and — most critically — credentials written to disk in plaintext. The component was rigid, heavy, and a security incident waiting to happen.

Designed and built a ground-up replacement in Python 3 that resolved every identified flaw. Added flexibility the predecessor lacked — letting customers supply their own validated queries against the read-only database instead of being locked to a static query. Shipped with its own automated test framework and a proper installer with logging. QA'd across 10 different *NIX distributions (it even ran on Solaris 10).

Cut the footprint from 50 MB+ to ~5 MB. Deployed to a major enterprise customer during beta with zero support tickets filed. Led the integration-engineering effort through to production and final documentation. The release became a Professional Services install package generating $50K+ in revenue.

A large enterprise needed to integrate more than 50 SaaS platforms with their privileged-access and secrets-management stack — a sprawling, high-stakes effort spanning identity providers, cloud platforms, and business systems, each with its own authentication model and API quirks.

Led a specialized development team through integrations with platforms including Okta, Microsoft Entra, Google Cloud, Amazon EC2, Workday, and Salesforce. Made the architectural calls on which OAuth2 grant type fit each case (client credentials, JWT assertion, and others), stood up isolated development environments per integration, and directed the team on the right REST and SOAP API approaches after learning each platform's API surface.

Delivered the full integration suite, hardening the client's identity and access posture and satisfying a compliance requirement that surfaced during review. The work kept the platform competitive and supported a significant long-term enterprise relationship.

Integrating security software with large, complex enterprise platforms — ServiceNow, Oracle Fusion, and others — for multiple large clients meant absorbing dense, unfamiliar API documentation and solving intricate integration problems under real delivery pressure.

Leveraged AI and structured prompt engineering as a force multiplier — accelerating research into unfamiliar APIs, drafting and validating integration logic, and working through complex tasks faster than traditional methods allowed. AI handled the grunt work; engineering judgment drove the decisions.

Completed large, complex integrations for enterprise customers on accelerated timelines, while keeping a human firmly in the loop on architecture and security decisions. A practical demonstration of AI used to ship real work, not as a buzzword.

A pharmacy ran a critical department on manual, error-prone processes, with sensitive health data that had to be handled under strict HIPAA requirements for security and retention.

Designed and built software that automated the department end to end: querying a live database, storing the data securely in the cloud, and backing up the on-premises fileshare — all architected to meet HIPAA standards for protected health information.

Replaced a fully manual workflow with a secure, automated, compliant system — reducing manual effort and error while keeping sensitive data protected and properly retained.