Async Apex at scale
Batch, Queueable and Schedulable jobs sized to governor limits, with retries, backoff and status you can see.
Async Apex, integrations, Lightning Web Components and data models that hold up in production. Platform Developer II certified, nearly four years on the platform.
From background jobs that process thousands of records to the screens people use all day.
Batch, Queueable and Schedulable jobs sized to governor limits, with retries, backoff and status you can see.
REST and SOAP callouts with OAuth 2.0, cached tokens, Named Credentials and polling jobs that recover when the other side fails.
Fast, accessible screens built with Lightning Data Service, GraphQL and SLDS, wired to Apex only where it earns its place.
Custom objects, relationships and sharing rules designed so the right people see the right records, and nothing more.
Bulk-safe triggers, Flows and calculation engines where every number has to add up, with tests that prove it.
Sandbox-first delivery with validated deploys, targeted test runs, code analysis and a rollback plan for every data change.
A batch that sends each file to an external AI service makes two callouts per record. Salesforce allows 100 callouts per transaction, but all of them must finish within 120 seconds. Move the slider and watch which limit breaks first.
Fits, with 20 seconds to spare on a slow day.
Assumes a slow day: 20 seconds of callout time per record.
Patterns from real production work, rebuilt from scratch with sample data. Every repo ships with Apex tests.
| File | Suggested type | Score |
|---|---|---|
| contract-signed.pdf | Contract | 96% |
| invoice-1042.pdf | Invoice | 93% |
| w9-form.pdf | Tax form | 90% |
| IMG_4471.jpg | Needs review | 54% |
| id-front.png | ID document | 88% |
Callout-enabled Batch Apex sends files to an AI classifier, and a Queueable files each one, with retry, backoff and an LWC review screen.
| Field | Was | Now |
|---|---|---|
| Opportunity.Amount | 48,000 | 52,500 |
| Account.BillingCity | Austin | Denver |
| Contact.Title | – | Added |
| Quote.Status | Draft | Draft |
OAuth 2.0 client-credentials calls with cached tokens, 401 and 429 handling, Queueable status polling and a versioned JSON diff shown in an LWC.
| Account | Debit | Credit |
|---|---|---|
| Cash | 254,218.75 | |
| Selling cost | 1,250.00 | |
| Receivable | 250,000.00 | |
| Premium | 2,500.00 | |
| Interest income | 2,968.75 | |
| Total | 255,468.75 | 255,468.75 |
Double-entry ledger lines that always balance, with day-count conventions, idempotent regeneration and variance routing to an approval queue.
Ownership followed through companies owned by other companies, with cycle detection, 100% checks at every layer and immutable approved snapshots.
The same route every time, so production never hears about a change first.
A requirement call, then the work split into small tickets.
Never in production. Code, metadata and permissions together.
Bulk-path tests, not one happy record.
So nobody else's work gets overwritten.
Validated with the tests that matter, deployed after sign-off.
Checked in the org itself, even when a deploy says it was cancelled.
Senior developer on a US client's Salesforce org: async processing, integrations, Lightning Web Components and data models, from requirement call to production release.
Built screening tools, data management screens, PDF generation and document automation on a production org.
Apex services, Lightning Web Components and two-way REST integrations on an enterprise Salesforce org.
A batch that keeps timing out, an integration that fails at 2am, a data model that outgrew itself. I'm always happy to talk it through.