Spandana Sunkara — Software Engineer

Software Engineer II

Crown Castle USA Inc — Canonsburg, PA

Apr 2024 – Feb 2026

Worked on migrating a legacy JSP monolith to React, GraphQL, and microservices. The old codebase was hard to change since everything was tangled together, so breaking it apart into smaller services made a real difference in how fast the team could move. Developer productivity went up around 40% once people weren't afraid to touch things.
Reworked Spring Boot backend workflows and added validation earlier in the deal processing pipeline so bad data got caught before it caused problems downstream. This cut out a lot of manual fix-up work the team was doing regularly and saved roughly 15+ hours a week across the board.
Replaced synchronous REST calls between services with Kafka-based event pipelines. The main motivation was that one slow service was taking down others, and decoupling them with async messaging fixed that. Service reliability improved by 30% under load, which was the clearest sign it was the right call.
Dug into slow API calls on the PostgreSQL and Oracle side, found N+1 issues, missing indexes, and some queries hitting the database repeatedly for the same data. Added caching where it made sense and fixed the queries. Latency dropped from around 800ms to 450ms, which was noticeable for users on the deal workflows.
Deployed services to AWS EKS using Docker and improved how services handled retries and errors when talking to each other. System stability went up 25% and we saw 15% fewer repeat incidents, mostly because failures were handled more gracefully instead of snowballing.
Investigated and resolved critical production incidents through log analysis and service debugging, which helped cut quarter-over-quarter incident recurrence by 15% and contributed to more consistent system stability over time.

React + GraphQL Spring Boot Kafka AWS EKS Docker + Kubernetes PostgreSQL + Oracle +40% developer productivity –15 hrs/week manual effort –30% cascading failures 800ms → 450ms latency +25% stability –15% incident recurrence

Software Engineer

Federal Soft Systems Inc — Bentonville, AR

Jan 2023 – Apr 2024

Built React components with Redux and TypeScript for internal reporting tools used by enterprise teams. Focused on making the interfaces actually usable and not just functional, so people could filter, navigate, and read the data without friction. Usability improved by 20% across the reporting workflows.
Wrote RESTful APIs in Java and Spring Boot to handle vendor data, things like fetching, updating, and validating records. Tightened up error handling and fixed some inconsistent endpoint behavior that was causing confusion on the client side. Defects reported by clients dropped by 25% after those fixes went out.
Helped set up CI/CD pipelines with Jenkins and Docker. Before that, deployments had a bunch of manual steps that were easy to mess up. Automating the process cut deployment time by 50% and made releases a lot less stressful for everyone on the team.

React / Redux / TypeScript Java + Spring Boot RESTful APIs Jenkins CI/CD Docker +20% usability –25% defects +50% deployment speed

Software Engineer

Plasticon Technologies — Hyderabad, India

Jul 2020 – Jul 2021

Built full-stack features using Python (Django/Flask) on the backend and JavaScript on the frontend across a few core products. Worked on reducing page load times by cutting unnecessary server-side processing and cleaning up how assets were loaded, which brought load times down by 20% and made the apps noticeably snappier.
Set up MongoDB replication and sharding to handle increasing data volume without the database becoming a bottleneck. Distributing reads across replicas and partitioning data across shards kept production running at 99.9% uptime even as usage grew.
Built Azure DevOps pipelines to standardize how the team released code. Everyone had their own way of doing deployments before, which made things inconsistent and slow. Once we had a common pipeline in place, release cycle time dropped by 40%.

Python / Django / Flask JavaScript MongoDB Azure DevOps –20% page load time 99.9% uptime –40% release cycle time

Samsung PRISM Developer

Samsung Research Institute — Bengaluru, India

Jan 2020 – Aug 2020

Part of Samsung's PRISM research program where I worked on building a real-time video classification model. Put together a custom dataset of 1,500 labeled videos, then tried out different architectures to get accuracy up while keeping inference fast enough to be practical. Ended up combining TSM with MobileNetV2, which hit 93% accuracy on the test set and was light enough to run in real time.

Deep Learning TensorFlow TSM + MobileNetV2 Video Classification Real-time Inference 93% accuracy 1,500-video dataset

Spandana
Sunkara

Where I've
Been & What I Built

What I Work
With Day-to-Day

Certifications &
Achievements

Looking for a
Mid-Level Engineer Role

SpandanaSunkara

Where I'veBeen & What I Built

What I WorkWith Day-to-Day

Certifications &Achievements

Looking for aMid-Level Engineer Role

Spandana
Sunkara

Where I've
Been & What I Built

What I Work
With Day-to-Day

Certifications &
Achievements

Looking for a
Mid-Level Engineer Role