Personal operating system

Building calm, useful software.

I’m Shiben, a software engineer in Bangalore. I like systems, thoughtful interfaces, photography, tennis, philosophy, and tools that make life feel more navigable.

BangaloreSoftware engineeriOS + systemsPhotographyMeditationSide projects
Shiben OS · working map
◈ SHIBEN OSclarity kernelsystemsattentionsoftwarewealthinputs: code, notebooks, maps, photos, money logsoutputs: useful tools, sharper choices

§0 · About

A short profile.

I’m drawn to systems that make messy real-world decisions clearer: where to live, what to build, how to debug, how to spend, and how to keep attention clean.

■ PROFILE

I’m a software engineer based in Bangalore. I like thinking from first principles, building side projects, understanding infrastructure, and documenting what I learn. The through-line is simple: make the invisible structure visible, then make better choices from it.

LocationBangalore
FocusSoftware systems, iOS, AI agents
InterestsPhotography, travel, philosophy
Current arcBuilding, learning, becoming sharper

§1 · Projects

Useful things, built as reasoning tools.

The projects I care about tend to start as personal friction, then become small operating systems for a specific decision.

Open projects
■ PROJECT

Name

NammaCircle

Concept / prototype
Stack
SwiftUI / Supabase / PostGIS / MapKit
Core idea
Help newcomers settle in Bangalore through locality intelligence, rent fairness, maps, Kannada learning, mentor support, and trusted local context.
iOS frontendSwiftUI

Native flows and app-state discipline

BackendSupabase

Auth, Postgres, Storage, Realtime

Spatial layerPostGIS

Locality intelligence and map queries

AI surfaceAgents

Reports, risk scoring, Kannada roleplay

■ PROJECT

Name

Personal Finance Tracker

Personal operating tool
Stack
Local-first app / automation / Gmail parsing / charts
Core idea
Track expenses, detect spending patterns, and decide whether major purchases fit the current financial plan before impulse takes the wheel.
Input layerGmail

Transaction alerts and receipt parsing

ModelLedger

Categories, months, merchants, amounts

OutputCharts

Trends, deltas, and purchase fit

PrincipleClarity

Money decisions made inspectable

■ PROJECT

Name

Monitor

Live
Stack
Next.js / Vercel / operational dashboard
Core idea
A focused operational view for keeping an eye on important signals without opening a full dashboard stack.
URLshiben.dev/monitor

Clean path under the main domain

ModeMonitor

For those obssessed with monitoring the situation.

SurfaceWeb

Available from the portfolio project index

PrincipleSignal

Make operational context visible

shiben.dev/monitor
■ PROJECT

Name

Countdown

Live
Stack
Next.js / Vercel / time utility
Core idea
Countdown for GTA 6 but later will have something more profound.
URLshiben.dev/countdown

Clean path under the main domain

ModeCountdown

< 200 days remaining!

SurfaceWeb

Linked from the project index

PrincipleClarity

Make time feel concrete

shiben.dev/countdown
■ PROJECT

Name

Rego Dojo

Live
Stack
React / OPA / Rego / Vercel
Core idea
A hands-on learning space for people new to Open Policy Agent, with guided Rego exercises and feedback loops that make policy authoring easier to practice.
AudienceOPA beginners

Built for new policy authors

Core skillRego

Practice rules, inputs, and evaluation

ModeDojo

Learn by solving small exercises

SurfaceWeb

Hosted as a standalone learning tool

rego-dojo.vercel.app

§2 · Engineering Notes

Research index for systems I keep returning to.

Notes are where the portfolio becomes a lab notebook: less about announcing expertise, more about preserving useful mental models.

Open notes
00Signal
01Question
02Trace
03Fix
04Learning
■ NOTE

Observability pipeline

Kubernetes → Prometheus → Cortex → Grafana / Logs → Loki / Traces → OpenTelemetry → Tempo

A stack is only useful when it helps you ask sharper questions during stress. Metrics explain shape, logs explain texture, traces explain path.

KubernetesPrometheusLokiTempoSLOs

TAKEAWAYI prefer systems where each tool has a clear boundary and the architecture can be reasoned about under stress.

■ NOTE

AI agents for infrastructure

Runbooks as tools, incident tickets as memory, deterministic exploration maps, and MCP servers as controlled action layers.

AI agentsMCPRunbooksIncident response

TAKEAWAYGood agents do not replace judgment. They compress the distance between question, evidence, and reversible action.

■ NOTE

Web development as one connected system

Frontend, backend, database, auth, deployment, observability, and product thinking are not separate rooms. They are pressure points in the same machine.

FrontendBackendAuthDeploymentProduct

§3 · Systems I Understand

A compact map of my technical surface area.

This is the part of the site that should feel scannable under pressure: labels, boundaries, and the pieces I can reason about.

Open systems
■ SYSTEM

Kubernetes

Pods, deployments, services, ingress, config, secrets, observability, security contexts, and Pod Security Standards.

clusters / workloads / networking / policy

■ SYSTEM

Observability

Metrics, logs, traces, dashboards, alerting, SLOs, incident debugging, Prometheus, Loki, Grafana, and OpenTelemetry.

signals / dashboards / incidents / SLOs

■ SYSTEM

AI agents

Tool boundaries, memory, deterministic workflows, evaluation, human approval loops, and agentic debugging surfaces.

tools / memory / evals / control

■ SYSTEM

iOS architecture

SwiftUI, navigation, app state, local models, Supabase clients, maps, and humane interaction details.

SwiftUI / state / maps / mobile UX

■ SYSTEM

Backend APIs

Auth, validation, database contracts, background work, integration edges, rate limits, and deployable service boundaries.

contracts / auth / jobs / reliability

■ SYSTEM

Databases

Schema design, indexes, RLS, migrations, analytics queries, Postgres extensions, and data modeling from product behavior.

Postgres / RLS / indexes / analytics

■ SYSTEM

Security

Secrets, least privilege, input boundaries, data access rules, transport assumptions, and operational blast radius.

trust / secrets / RLS / blast radius

■ SYSTEM

Personal finance automation

Expense ledgers, transaction parsing, category drift, compounding habits, and software that makes tradeoffs visible.

ledger / parsing / habits / wealth

§4 · Photography / Travel

Attention outside the terminal.

Photography is how I practice noticing. It softens the engineering brain without turning it off.

Open field notes
■ FIELD NOTE

birds / water / morning light

Bangalore lakes

Birds, morning light, urban nature, and the slow discipline of noticing what the city hides in plain sight.

■ FIELD NOTE

history / games / landscapes

Japan / Tsushima

A long-running dream shaped by history, games, landscapes, and the atmosphere of places that feel almost mythic.

■ FIELD NOTE

streets / cafes / attention

Urban walks

Cafes, streets, weather, architecture, and the small texture of city life when you are not rushing through it.

§5 · Experiments

Live lab, not finished identity.

A few parts of life are better treated as experiments: define the signal, observe honestly, adjust the system.

Open experiments
■ EXPERIMENT

compound rate

Building wealth

Reducing unnecessary expenses, increasing income, investing consistently, and using software to make better money decisions.

■ EXPERIMENT

attention quality

Mental clarity

Meditation, attention, emotional resilience, and learning how to act from a higher mental plane without becoming abstract.

■ EXPERIMENT

energy baseline

Fitness

Strength, energy, sports, and sustainable physical confidence that supports the rest of the system.

■ EXPERIMENT

study surface

Learning

Japanese, art, systems, and deep technical understanding treated as a long, compounding notebook.

§6 · Contact

Open to thoughtful conversations.

Software, systems, AI agents, infrastructure, photography, side projects, and the occasional first-principles detour.

■ OPEN CHANNELS

I like conversations where curiosity is concrete: what are we trying to understand, what evidence do we have, and what small thing can we build next?