all experience
completed[Sept 2023 – Oct 2024]

Innovative IT

Software Developer · including 2.5 months parttime

  • Performance: Owned bottleneck diagnosis and resolution through indexing and queryset optimization, driving measurable reductions in API response times across client applications.
  • API Design: Architected and owned production-ready REST APIs with DRF, enforcing secure auth patterns and optimized serialization logic for low-latency data delivery.
  • Async Architecture: Designed and owned Celery and Redis pipelines to offload reporting workflows and background processing tasks, directly improving frontend responsiveness.
  • Backend Engineering: Designed relational schemas and Django services built for data integrity and scalability across web and mobile surfaces.

The story

idea → framing → build → deploy → what's next

  1. 01 IDEA

    Performance: Owned bottleneck diagnosis and resolution through indexin

    Innovative IT was my first full-time engineering role after the MCA coursework. It was a small consulting shop where everyone wore multiple hats. My focus was the Django and DRF backend that powered three client-facing products: two web apps and a mobile app, all hitting the same REST surface.

  2. 02 FRAMING

    The first big project was a customer-portal API for a logistics client.

    The first big project was a customer-portal API for a logistics client. I owned the schema end to end: relational modeling, DRF serializer shapes, and the authentication strategy. We shipped v1 with token-based auth, pagination, filtering, and a handful of specialized endpoints for shipment tracking. The mobile client was the harder constraint. Low latency, strict freshness. I structured the read path to use materialized views for the most-queried rollups, and the write path to publish change events to a lightweight WebSocket layer.

  3. 03 BUILD

    Performance work kept showing up.

    Performance work kept showing up. I'd inherited a couple of hot endpoints running at 800ms+ median latency. The fix was always the same shape: find lazy-loading patterns, replace them with select_related and prefetch_related, add the missing composite indexes. One endpoint dropped from 1.2s to 80ms after a single migration. The discipline stuck. Every endpoint I wrote from then on started with "what's the slowest query path here, and is the index supporting it?"

  4. 04 DEPLOY

    Deployment turned out to be where I learned the most.

    Deployment turned out to be where I learned the most. The first v1 shipped on a bare EC2 with no automation. By month six I had Fabric scripts for app, migrations, collectstatic, and restart, then Ansible roles for the broker and the DB. The blue/green pattern (ALB swaps target groups with no downtime) came directly from this, and it's the same pattern I codified at NexBell. The lesson is that deployment is a forcing function for the rest of the architecture.

  5. 05 WHAT'S NEXT

    Outcome — Performance: Owned bottleneck diagnosis and resolution throu

    What I learned at Innovative IT. Write tests before the second PR. Keep serializers thin (one shape, one place to change). Choose the right indexes up front rather than tuning late. The DrishtiAI pipeline and Algocode both trace patterns I honed here, even though neither uses Django. What's next, if the consulting path ever opens again. I'd add observability (Prometheus and Grafana on day one) and a real per-tenant data model so each client could ship independently without shared-schema risk.

Keywords

4 techs · links go to /stack
DjangoDRFPostgreSQLREST APIs

Related projects

1 project shipped alongside work at Innovative IT
shipped★ 8

ImgTwist

Image showcase platform. Django, async processing, social features.

DjangoCeleryRedisPostgreSQL

Related writing

6 posts reference Innovative IT

Product case study

For the architecture deep-dive: three-layer build, isolation strategy, deployment topology. Read the case study: ImgTwist on /work.