HarshitRaj
Curiosity-led builder. Calm execution, consistency, and the kind of growth that lasts.
Harshit Raj
I’m a curiosity-led builder. I like understanding how things work, then turning that understanding into something real: a system, a tool, a habit, or a craft.
I’ve built and shipped systems at scale at Amazon and AdPushup, but I don’t want this site to feel like a resume. It’s more of a snapshot of what I’m into: consistency (gym/running), perspective (travel/mountains), and clarity (writing).
The thread connecting everything is simple: show up, iterate, learn, and keep it honest — no heroics, just steady progress.
How I got here
I’m drawn to how things work — and the discipline it takes to get better at anything.
I didn’t start coding because I wanted to build startups or chase titles. I started because I liked understanding how things actually work under the hood. In college, I was more interested in breaking programs, debugging weird issues, and figuring out why something failed rather than just getting output.
Early on, I got exposure to real production systems at scale — the kind that teach you humility fast. I learned that tiny mistakes become real outages, and that good logging, observability, and ownership matter more than cleverness.
I’ve shipped at scale at Amazon and AdPushup, but I don’t want to turn this page into a list of job bullets. What matters to me is the shape of the work: reliability, calm execution, and making things simpler and more predictable over time.
Outside work, I train. I run (even when I don’t feel like it). I travel to reset perspective. I write to think. Those habits keep me grounded — and they show up in how I build, learn, and show up every day.
Tools I trust in production
Not a keyword list — this is what I reach for, and why, when the goal is calm systems.
Backend & APIs
- Java + Spring Boot
Production APIs, strong conventions, and boring reliability at scale.
- REST (and pragmatic contracts)
Clear boundaries, versioning discipline, and predictable operability.
- Go (selectively)
Low-latency services, concurrency-heavy workloads, and simpler binaries.
Messaging & Streaming
- Kafka
Durable event pipelines where throughput and replayability matter.
- Kafka Streams
Real-time transforms/anomaly detection without building a new platform.
Data Stores
- SQL (Postgres/MySQL)
Core source-of-truth data with well-modeled schemas and constraints.
- Couchbase
Document + caching patterns where read latency and flexibility matter.
Observability
- Prometheus + Grafana
Metrics-first debugging, SLO visibility, and honest capacity signals.
- Alerting hygiene (PagerDuty)
Fewer pages, higher signal, and faster incident triage.
- Structured logging
Searchable, consistent logs that actually help during incidents.
Cloud & Infrastructure
- AWS (Lambda, CloudWatch)
Managed primitives with clear ownership and cost boundaries.
- Docker
Reproducible builds and consistent deploys across environments.
- IaC (CDK/Terraform)
Make infra reviewable, repeatable, and enforceable over time.
Engineering workflow
- CI/CD (Jenkins)
Automate boring steps; push confidence left with tests and checks.
- Testing (integration + e2e where needed)
Catch regression at boundaries that break in production.
- Design reviews
Decisions age — writing them down keeps systems coherent.
The shape of my work
I’m intentionally not listing titles/dates here. This is the kind of work I enjoy: systems thinking, reliability, and clear ownership.
- Core backend flows (think: high-stakes state transitions, retries, and edge cases).
- Operational improvements: calmer on-calls, clearer alerts, and fewer surprises.
- Cost + performance trade-offs: making systems predictable without over-optimizing prematurely.
- Event-driven pipelines and ingestion flows (throughput, retries, backpressure, and correctness).
- Monitoring + debugging loops: making issues easier to see and faster to fix.
- Real-time signals: catching anomalies and traffic shifts without drowning in noise.
Contributions & Maintenance
Small, targeted changes that improve correctness, UX, and long-term maintainability.
Fractional-offset scheduling, UTC parsing refactor, and docs improvements.
Added unit tests and CI coverage for spinner components (Go).
UX fixes and stability improvements for the v1.0 release (Go + HTMX).