04 / LAB · QUESTIONS WORTH EXPLORING

Where ideas
get stressed.

A technical shelf for the ideas that keep appearing across my builds — concurrency, graphs, service boundaries and algorithms.

AKARSH / EXPERIMENTAL LIVE
LAB
STATELOADGRAPHBOUNDARY
VARIABLECOMPLEXITYMODEOBSERVE → ITERATE
04 / 05
01 / DISTRIBUTED

Shared state
under load.

Rate limiting becomes interesting when several application instances must agree on the same truth. The experiment focuses on atomicity, refill semantics, retry behavior and what Redis changes about the design.

JAVAREDISLUA
20 TOKENS
ATOMIC CHECK / CONSUME / REFILL
02 / INTELLIGENCE

Graphs +
anomalies.

Instead of treating transactions as isolated rows, model them as relationships. Connected components, cycles and anomaly scoring create a different lens for suspicious behavior.

NETWORKXMLPLOTLY
SUBGRAPH / RISK CLUSTER DETECTED
03 / ARCHITECTURE

Boundaries
that matter.

Service decomposition is useful only when the boundaries create clarity. I explore how authentication, account state and payments can remain independently understandable while still composing into one product.

NODESPRINGPOSTGRES
AUTH
ACCOUNT
PAYMENT
POSTGRESQL
BOUNDARY / CONTRACT / PERSISTENCE

ALGORITHM PRACTICE

Keep the
edge sharp.

Problem solving and constraint thinking stay in the loop because good system design still depends on fundamentals.

01   graph[u].push(v)02   if (seen[v]) cycle()03   score += anomaly(x)04   ship();
LC
OPEN AKARSH2905 ↗