Engineering Escalation Engine
Next.js and TypeScript system for structured intake, AI-assisted analysis, deterministic escalation scoring, evidence validation, human approval and auditable Engineering handoffs.
Project write-up coming soon →Evidence first
95%
Issue reproduction rate
Turning customer reports into reproducible technical evidence before Engineering picks them up.
30%
Less Engineering investigation time
Better reproduction, logs and context mean less time spent reconstructing the problem from scratch.
25%
Faster P1 and P2 resolution
Observability dashboards and clearer escalation paths give critical incidents better context earlier.
Where I go deep
01 · Investigation
Technical investigation and reproduction
I work from the customer report back toward the system. That means reproducing the issue, checking logs, testing API behaviour, narrowing conditions and documenting what is actually known.
REST APIsPostmanDatadogKibanaElasticsearchBrowser DevToolsDNS
What this looks like in practice: 95% issue reproduction rate and 30% less Engineering investigation time.
02 · Identity and integrations
Authentication, SSO and enterprise integrations
A large part of enterprise support is understanding where systems meet. I troubleshoot identity, provisioning, API and integration problems across customer environments rather than treating them as isolated tickets.
SAMLOAuthSSOSCIMSlackMicrosoft TeamsSMTP
What this looks like in practice: complex enterprise migrations, authentication failures and integration troubleshooting across 100+ accounts.
03 · Escalation systems
Incidents, escalation and Engineering handoffs
I care about the point where Support hands a problem to Engineering. The quality of that handoff changes how quickly the issue can move, so I focus on evidence, reproducibility, severity, environment and what has already been ruled out.
RootlyJiraConfluenceDatadogP1 and P2Escalation playbooks
What this looks like in practice: structured P1 and P2 escalation ownership and a portfolio project built around evidence-based Engineering handoffs.
04 · Support operations
Workflows, knowledge and customer signals
I look for the repeated work behind the queue. If the same issue, request or escalation keeps appearing, I want to understand whether the fix is better documentation, a workflow, automation or a clearer path into Product and Engineering.
SalesforceFreshdeskZendeskJiraKnowledge workflowsFeedback intake
What this looks like in practice: customer feedback workflows, escalation playbooks, internal guides and repeatable support processes.
05 · AI and automation
AI-assisted support systems
I use AI where it helps structure, analyse or draft support work, but I keep evidence and decision authority separate from the model. The interesting part for me is not adding AI to a workflow. It is deciding where AI belongs and where it does not.
OpenAI Responses APIAnthropic Claude APITypeScriptJavaScriptZodCloudflare WorkersSupabase
What this looks like in practice: deterministic validation, human approval points, grounded prompts and secure API routing across my portfolio projects.
06 · Building
Enough software depth to build the workflow
I am not trying to position myself as a software engineer. I use code because sometimes the clearest way to explore a support problem is to build the system I think should exist.
Next.jsTypeScriptJavaScriptSQLGit and GitHubDockerHTML and CSS
What this looks like in practice: shipped AI tools, REST APIs, persistent data models, tests and production deployments.
Proof through projects
Feedback Loop
Turns scattered feature requests into grouped themes, structured product signals and a product brief without losing the original customer evidence.
View case study →Support Ticket Analyser
Matches tickets against existing knowledge, identifies documentation gaps and uses AI to draft reviewable KB content when the gap is real.
View case study →Technical toolkit
InvestigationDatadog, Kibana, Elasticsearch, Browser DevTools, Postman, log analysis, issue reproduction
APIs and identityREST APIs, SAML, OAuth, SSO, SCIM, DNS, SMTP, authentication troubleshooting
Support operationsFreshdesk, Salesforce, Zendesk, Intercom, Jira, Confluence, Rootly, Slack
Build and automationNext.js, TypeScript, JavaScript, SQL, Git and GitHub, Docker, Zod, Supabase, Cloudflare Workers
AIOpenAI Responses API, Anthropic Claude API, prompt design, structured outputs, evidence validation, AI workflow design
Learning and certifications
IBM AI FundamentalsGenerative AI: Prompt EngineeringDatadog: Performance MonitoringREST APIsSQLLinuxPython, ongoing