Skip to content
Mirai Minds

Shipped work: AI systems in production

In short

This is the work Mirai Minds has shipped: AI systems built for clients, and platforms we run ourselves. Each case study covers the problem, how the system works, the models and stack behind it, and what changed, with the source of every number. Startup clients are named; enterprise and public-sector clients are anonymised.

Shipped to productionUpdated

Shipped to production, updated 23 September 2026
ProjectWhat we builtResultYearStatus
Resin manufacturer (India)Catalog chat assistant with text and photo product searchPilot on a 15-product catalog sample2026–present Pilot
Photo ExperienceGenerative photo editing, face search and the image pipeline18.7 s per generative photo on one A100 (was 60–90 s)2023–present Pilot
iKoMatchWhatsApp and voice fundraising assistant with explainable matching6,717 investor records; each day's matches graded by an LLM judge2026 In production
CallAuditAIIn-houseCall audit service: recordings scored against a weighted rubricScorecards for an admissions counseling team's recorded calls2026 In production
Ethnic-wear e-commerce brand (India)Visual product search and abandoned-cart recovery callsVisual search live for store staff since July 20262026 In production
WhatsApp AgentsIn-houseWhatsApp agent service with memory, summaries and document searchRuns iKoMatch's founder conversations on WhatsApp2026 In production
AI outbound-calling startupPipecat voice bot, campaign dialer and per-call recordingLive for outbound campaigns2025–2026 In production
Voice AgentsIn-houseMulti-tenant platform for AI phone agents and calling campaignsLive at voice-agents.miraiminds.co2025–2026 In production
Indian public-sector refineryDocument AI: engineering drawings to SAP-ready bills of materialDelivered and handed over; every row reviewed before it reaches SAP2025–2026 Delivered
US logistics software companyChat agent over the client's platform APIs through MCPDelivered; every tool call's arguments are checked before it runs2025 Delivered
Gurbani learning appTransliteration, explanation and speech API for Gurbani linesDelivered to the app team as an authenticated API2025 Delivered
Myva.ai and CallPaaSDocument-grounded chatbot platform and call-center AI with GPTBoth products built and extended across 2023–20252023–2025 Delivered
RouhDesktop computer-use agent with MCP tools, scheduling and memoryWorking desktop prototype, reviewed by the founder2024–2025 Prototype

Agents that read, decide and act inside your tools, with every tool call logged and a person approving the risky ones.

Phone agents that answer, qualify and follow up in English or Hindi, and bring in a person when the caller needs one.

Language models grounded in your documents and data, measured on your own test set, with people reviewing what matters.

Vision models that find faces and products, cut out backgrounds and edit photos, fast enough for a queue of real customers.

Products we build and run ourselves

How we measure and what we leave out

Every entry in the table above is a system that exists: live in production, delivered and handed over, in pilot, or a working prototype. The status column says which.

Case studies follow the same order: the problem, what we built, how it works in production, and what we would change. That last part is deliberate. Every system has trade-offs, and knowing them is more useful to you than a clean story.

Asked on the first call

How are the numbers in these case studies measured?

Each metric names its source next to it: a benchmark run, a load test, a production log or the client's own records, with the month it was measured. If we could not measure something, we describe the outcome instead of estimating a number.

Why are some clients anonymised?

Enterprise and public-sector clients often can't be named publicly. We describe them by industry and country instead, and leave out anything that would identify them, such as locations, partner names or document samples.

Can we speak to a client you worked with?

Where the client agrees, yes. Ask on the first call and we will check who is happy to talk about the project.

Have a system in mind? Let's scope it.

A 30-minute call with an engineer who has shipped this before. You leave with a plan, a rough timeline and what it would take — whether or not we build it.