What we build, fix and take responsibility for.
Nine practice areas. Each one is work we have shipped and then operated — not a capability slide. If a request falls outside them, we will tell you plainly rather than learn on your budget.
Most engagements start as a two-week advisory sprint or a fixed-scope project. Send us the problem and we will come back with a written scope.
Go backend engineering
Services that stay up, stay readable and stay cheap to run. Go is our primary language and has been the production stack for the last three years.
APIs and microservices
REST services in Go and Gin with a designed data model, sensible error surfaces, and request handling that does not fall apart at the 95th percentile.
Concurrency and queues
Goroutines and channels used deliberately — ordered, back-pressured processing for ingestion and generation workloads, rather than unbounded fan-out that melts the database.
Schema and query design
PostgreSQL schemas that will still make sense in two years, indexes chosen from real query plans, and Redis used for what Redis is good at.
Rescue and refactor
Inheriting a service nobody understands: reading it, documenting what it really does, adding the tests that matter, then changing it safely.
Cloud architecture & migration
Getting workloads onto a platform that scales down as well as up — and moving whole estates between providers without the business noticing.
Serverless and containers
GCP Cloud Run and Cloud SQL, AWS equivalents, Cloudflare Workers at the edge. We have migrated a single-VM service to serverless and watched the bill and the pager both go quiet.
Provider-to-provider migration
Inventory, dependency map, cutover sequence, rollback plan. Executed once already across a full estate, moving billing currency and data residency at the same time.
Deployment and operations
CI/CD, secrets handling, environment separation, automated backups, monitoring and a documented incident path — set up so your team can run it after we leave.
Linux and Windows estates
Ubuntu and Windows Server administration, hardening, patching regimes and the unglamorous cleanup that keeps an estate from drifting.
Cloud cost & FinOps review
A fixed-fee teardown of your cloud bill, ending in a ranked list of changes with the rupee value and the risk written next to each one.
Where the money actually goes
Spend attributed to services, environments and teams — including the idle estate, over-provisioned instances and egress nobody budgeted for.
Architecture-level savings
The big wins are rarely a smaller instance. Moving always-on VMs to request-scaled services, fixing a chatty integration, or re-platforming entirely.
Currency and vendor terms
For Indian businesses billed in USD, moving to an INR-billed provider can be worth more than any right-sizing. We have executed that move end to end.
An honest verdict
If your estate is already lean, the report will say so in the first paragraph. We would rather lose the follow-on work than manufacture it.
AI engineering & automation
Language models applied where they beat the alternative — and honestly costed, including the cases where a regular expression would do the job better.
Document and data extraction
Turning invoices, statements, contracts and semi-structured feeds into records your systems can use, with confidence scores and a human review path.
Assistants over your own systems
Retrieval over your documents and databases, wired to real APIs with real permissions — not a chatbot that hallucinates your policies.
Agentic automation
Multi-step workflows that read, decide and act against your systems, with logging, spend caps and a stop button. Built on the Claude and Gemini APIs.
Enabling your engineers
We have used AI-assisted development in a production team's workflow since April 2025. We will set up the tooling, the review discipline and the boundaries that keep quality up.
Enterprise applications & ERP
Fourteen years around PeopleSoft Financials, nine of them supporting a global bank's finance estate. This is depth you cannot hire quickly.
PeopleSoft FSCM functional support
General Ledger, Asset Management, Receivables, Billing, Lease Administration, AP and Travel & Expenses — incident, change and release governance included.
Upgrades and implementations
9.1 upgrade analysis and technical lead work on General Ledger and allocations, for banking and retail clients.
Reporting and interfaces
nVision, PS Query, SQR reports, App Engine programs and Integration Broker services — plus bolt-on interfaces to everything the ERP has to talk to.
Automation of the tedious parts
PeopleSoft and Python automations that took recurring finance processes off people's desks entirely.
Data integration & pipelines
Bringing in data from sources that were never designed to agree, and making the result something you can reconcile and defend.
Multi-source ingestion
Feeds, files and APIs from many partners, normalised into one model — with schema drift detected rather than silently absorbed.
Attribution and matching
Getting the right record to the right party. We built commission attribution across several affiliate networks while keeping affiliate identity hidden from the sources.
Reconciliation and replay
Runs that can be re-run safely, totals that tie back, and a clear answer when someone asks why last Tuesday's numbers changed.
Backup and recovery automation
Python automations that package remote servers into verified archives, on a schedule, with restores actually tested.
Analytics & reporting platforms
Getting the numbers out of the system and in front of the people who act on them.
Metabase, set up properly
Modelled sources, sane permissions, dashboards people actually open — instead of a tool installed once and abandoned.
Custom analytics engines
When the question is not a dashboard question. We built an engine that surfaced winning products to the affiliates selling them, which no generic BI tool would have answered.
Operational reporting
The recurring reports finance and operations need on a fixed day, produced automatically and correctly, every time.
Governance & audit readiness
Nine years owning SOX controls inside a global bank and answering to Big-4 auditors. We know what evidence holds up and what gets a finding.
Control design and evidence
Access, change and release controls defined so they can actually be operated — then evidenced in a way that survives testing.
Audit liaison
Walkthroughs, sample requests and remediation plans, handled by someone who has sat on that side of the table for years.
SDLC and release governance
A change process proportionate to your size — enough control to pass review, not so much that nothing ships.
Vendor, SOW and exit management
Holding suppliers to the statement of work, and running the handover when a contract ends — including the full code-and-data transfer we have done ourselves.
Fractional leadership & team building
For companies whose engineering has outgrown improvisation but does not yet justify a full-time technology head.
Building the function
We built an 11-person engineering team from an early stage — full-stack, front-end, operations and DevOps — covering hiring, onboarding and performance management.
Standards and review
Coding standards, review discipline and architecture decisions that stick, set by someone who still writes and reviews the code.
Growing juniors
Turning fresh graduates into engineers who deliver independently. This is slow, deliberate work, and it is the cheapest capacity you will ever add.
Accountability to the business
Translating between engineering and the board: what it will cost, when it will land, and what the risk actually is.
Bring us the awkward one.
The service nobody wants to touch, the bill nobody can explain, the integration that breaks every quarter. Describe it in a few lines and we will come back with how we would approach it.