// work — 7 case studies
Each write-up covers the constraint, the architecture, what shipped, and the numbers — plus what I'd do differently.
RaghuvarTech · open source
The full LGTM stack plus profiling and alerting, packaged as a decoupled sidecar that never starves the applications it watches, with distributed tracing across FastAPI and Next.js.
logs · traces · metrics · profiles
signals
memory-capped
host impact
RaghuvarTech
Replaced click-ops and shared credentials with a single IaC repository: reusable Terraform modules, Ansible roles, plan-before-apply pipelines, OIDC auth and nightly drift detection.
3
environments
0
long-lived cloud keys
CaseCloud (casecloud.in)
A case-management platform for law firms where every tenant is isolated at the database layer and every AI answer cites the judgment it came from.
80M+
judgments indexed
6K+
legal templates grounded
InkHub.in / BRMH.ai
An AI platform that turns a prompt into a working e-commerce app (APIs, schema, frontend), re-architected from a monolith into event-driven containers with a zero-downtime migration and a 76% lower cloud bill.
−76%
monthly cloud spend
migration downtime
Vedico Infrastructure
A NOC dashboard that watches hundreds of field technicians live, flags theft and proximity anomalies, and escalates SLA breaches automatically through a four-tier chain.
sub-ms
broadcast latency
50 m
geofence radius
Independent product
A desktop copilot where LLM agents navigate real immigration portals, extract applicant data from messy documents, and fill long forms like I-485 and DS-160 with the human kept in the loop.
90%
manual entry reduced
I-485 · DS-160
forms supported
EatTogether & Co. (Seattle, remote)
Replaced a static price table with a realtime engine that personalises rates from availability, spend history and location, and pulled the pricing path out of a monolith into independently deployable services.
<100 ms
quote latency
signals per quote
// need something similar?
Pick a time that fits you, or just