DATA & PERFORMANCE
PostgreSQL Dashboard Optimization: 15s → 2s
Moving repeated reporting work out of the request path with materialized views, pre-aggregation and revised queries.
Node.js · PostgreSQL · PrismaRead case study
ALI ANJUM / ENGINEERING WITH OWNERSHIP
Senior Backend & Applied AI Engineer
I solve difficult backend, data and integration problems — and turn AI prototypes into reliable production systems.
Node.js · TypeScript · PostgreSQL · MongoDB · AI systems
- 10+ years software development
- Toptal-vetted
- Technical Team Lead experience
- Italy · Remote
Dashboard optimization with materialized views, pre-aggregation and revised queries.
Explore the case studyNODE.JS / AWS5 min → 15 secSequential browser-heavy processing replaced with concurrent Axios/Cheerio work.
Explore the case studyMONGODB26.6M records~1s chart queriesPre-aggregated time-series data and chart-specific queries.
Explore the case studyResults from separate projects and workloads. Each case study explains the measurement context.
01 / PROBLEMS I SOLVE
Recognize the problem?
Start with what is getting in the way.
01
Backend is slow
Investigate API, query, concurrency and architecture bottlenecks and implement measurable performance improvements.
See relevant evidence02Data has outgrown the application
PostgreSQL/MongoDB analysis, aggregation, materialized views, reporting architecture and pre-computation.
See relevant evidence03AI prototype needs to become a product
Model integration, backend APIs, validation, fallbacks, observability, data flows and production architecture.
See relevant evidence04Integrations keep breaking
Payments, CRM systems, webhooks, queues and third-party APIs.
See relevant evidence05Difficult production issue has no clear owner
Reproduce → measure → investigate → isolate → implement → verify.
See relevant evidence02 / SELECTED CASE STUDIES
The reasoning behind the result.
PRODUCTION DEBUGGING
When AI Was Confidently Wrong: Debugging an Apple Pay Production Failure
An Apple Pay failure showed why a plausible AI answer must still survive contact with observed browser behavior.
Stripe · Apple Pay · Browser user activationRead case study
03 / HOW I WORK
Evidence before assumptions.
- 01Understand
- 02Measure
- 03Reproduce
- 04Isolate
- 05Fix
- 06Verify
Understand the business impact, reproduce the issue, identify the real bottleneck, implement the smallest safe change, measure again and monitor the result.
04 / APPLIED AI & EXPERIMENTS
Keep building. Keep learning.
Travel Junior
Production AI application architecture
Svelte and TypeScript with Fastify, Gemini, Google Places and PostgreSQL. Caching, validation and fallbacks, metrics, and PDF export. The repository is private.
Explore Travel JuniorFocusFeed
Personalized discovery
Gemini embeddings, ranking and personalization with PostgreSQL/pgvector and Supabase. No public usage or adoption claim.
View project details05 / EXPERIENCE & OWNERSHIP
From hands-on developer
to technical team lead.
At a Toptal client, I worked directly with the technical CEO on backend systems, AI lead qualification, payments, HubSpot and difficult production issues.
Software development since 2014, across independent client work and engineering teams.
CLIENT EVIDENCE
“Muhammad did a great job. He improved on the previous NodeJS code base and made it run nearly 20X faster! He was readily available for any small changes we requested and understood immediately what we're asking for. Would recommend.”
06 / ASK ALI
A shorter route to the right evidence.
Ask my AI assistant about a case study, my experience, or a technical problem. It can point you to relevant work. The opinions—and occasional stubborn debugging sessions—are mine. The assistant is AI.
Ask Ali a question
LET’S SOLVE IT
Have a difficult engineering problem?
If you’re dealing with a backend bottleneck, production issue, unreliable integration or an AI prototype that needs production engineering, let’s discuss it.
Discuss the problem