I build and scale backend systems with a focus on performance, reliability, and clarity of design.
I’m currently pursuing my Master’s in Computer Science at Penn State, and previously worked for 2 years at Enphase Energy, where I engineered and optimized cloud services used by thousands of users.

I enjoy the kind of problems that require understanding data flow, concurrency, and real system behavior—not just syntax.

• Designing and scaling distributed and backend systems
• Diagnosing and resolving performance bottlenecks
• Building clean, maintainable system architectures
• Teaching and explaining complex concepts simply and intuitively

A multi-threaded MapReduce framework built in C++ using gRPC for distributed execution.
Resolved reducer-side lock contention by introducing double-buffered batching + fine-grained synchronization, achieving 6–7× throughput improvement under large workloads.

Developed a Random Forest–based triage model that improved fairness and consistency compared to manual prioritization.
Reduced subjective decision bias and supported more transparent resource allocation.

• Cloud-native service design (AWS + containerized microservices)
• Distributed coordination + consistency patterns
• Full-stack systems with real user interaction, not just demos

I like systems that are:

• Understandable – simple where possible, structured where needed
• Observable – you should know why something works, not just that it works
• Sustainable – scalable systems stay clean only with deliberate design

I also serve as a Graduate Learning Assistant, mentoring students in C++ and algorithmic reasoning — teaching keeps me honest:

If I can’t explain it clearly, I don’t understand it well enough.

• Email: kinshuk048@gmail.com
• LinkedIn: https://www.linkedin.com/in/pratyushkumar123

I play competitive badminton 🏸 — it keeps me sharp, disciplined, and team-focused.
Also: I enjoy helping others debug the “impossible bugs” — it’s my version of puzzles.