# Saasi — Custom software, shipped like product.

## What we do
Saasi is a software and SaaS engineering studio for teams that need product-grade
engineering on contract. We are AI-native: a large part of our work is building agentic
systems — multi-agent LLM pipelines, RAG, and MCP (Model Context Protocol) servers that
make products usable by AI agents — alongside SaaS products, Shopify apps, and full
platforms. We design, build, and ship end to end (architecture, backend, web, mobile,
infrastructure, launch) and hand over clean. We are not a staffing agency and not a
throwaway-MVP shop; we work like a senior product team that owns the outcome.

Typical engagements: building an AI-native or agentic product from zero to launch; making
an existing product agent-ready (MCP servers + tool APIs); building embedded Shopify apps
and custom storefronts; standing up hard platform capabilities (multi-tenancy, billing,
real-time, search, data pipelines); rescuing or re-architecting a stalled codebase; or
embedding as a fractional senior engineering team.

## How we work
- Scope — map the problem, constraints, and success criteria before writing code. A clear plan, milestones, and a fixed shape of work.
- Ship — build in production-grade increments: typed, tested, reviewed, deployed. Working software early and often, never a big-bang reveal.
- Scale — harden, document, and hand off, or stay on to keep building. The code is yours.

## Services
- End-to-end product build — 0 to 1 through launch: product architecture, backend, web, mobile, and cloud infrastructure, owned end to end.
- Agentic AI & LLM — multi-agent systems, LLM tools, and RAG shipped to production with evals and guardrails.
- MCP & agent-ready APIs — MCP servers, tool APIs, OAuth scopes, and Schema.org so AI assistants can act on your product.
- Shopify apps & storefronts — embedded Shopify apps and custom storefronts, including agentic commerce that AI shoppers can browse and buy from.
- Platform capabilities — multi-tenancy, billing, auth, real-time, search, and data pipelines.
- Codebase rescue — stabilize, re-architect, and get shipping again.
- Fractional senior team — embedded senior engineers for a fixed term.
- Architecture & due diligence — technical review, architecture, and an executable delivery plan.

## Selected work
- Tarasvin (logistics / mobility) — a shared driver-pool platform running day-prior scheduled B2B routes and real-time parcel and ride dispatch on one supply pool, with a public REST/OpenAPI API and an MCP server for agent-driven booking. 30+ Go and Python microservices, four web portals, and driver + customer mobile apps. Delivered production-grade and launch-ready. Live at https://tarasvin.app
- Vanigar (fintech / investing) — AI-native equity research: a multi-agent LLM desk (LangGraph) reads filings, charts, and news across the NIFTY, debates bull vs. bear, and returns a BUY/HOLD/SELL verdict with its reasoning. Shipped across web, admin, and native iOS/Android, with RAG over Elasticsearch. Live at https://vanigar.app

## Stack
TypeScript, React, Next.js, and Node; Go and Python services; Postgres; Expo (React Native)
for mobile; LLM and agent tooling (LangGraph, RAG, MCP); cloud-native infrastructure
(AWS/GCP), containerized and observable. We pick boring, proven tools by default and
reserve novelty for where it earns its place.

## Engagement model
Fixed-scope projects or monthly retainers, milestone-based with clear deliverables. All
code, infrastructure, and documentation are owned by the client.

## Voice
Plain, not clever. Confident, not boastful. We write for the engineer in the room — verbs
over jargon, "ship" over "solutions and synergies."

## Contact
Email: enquiries@saasi.work
Web: https://saasi.work
Book a scoping call: https://saasi.work/book