FBN
Exhibit III / 16Full-Stack — 2026Full-stack engineer, sole build← Index

AssetStream

An operating system for equipment finance: leases, IoT-metered billing, risk scores and AI-guided remarketing, under one roof.

AssetStream

FBN.2026.03

AssetStream

Full-stack engineer, sole build

Django 5, Django REST Framework, Celery, Redis, 2026

1

command to raise the entire stack

8

backend domains, from originations to remarketing

3

roles served: admin, analyst and lessee

6

containers orchestrated in concert

The problem

Enterprise asset finance still runs on opaque legacy systems: leases in one place, usage data in another, and the value of a returning machine decided by instinct.

The approach

A Django REST core organises the business into originations, servicing, payments, remarketing and workflow domains. Celery workers on Redis ingest equipment telemetry and close invoices every cycle; a scikit-learn regression prices each machine at the end of its lease; an AI engine on Groq's Qwen3-32B scores risk, anticipates maintenance and answers analysts in plain language. A Next.js 15 dashboard fronts the whole estate, and the full stack rises with a single docker compose command.

The outcome

The complete asset-finance lifecycle, from origination to remarketing, running on Render in production and reproducible on any laptop in one command.

What it does

4 capabilities

The lease, end to end

Origination, servicing and end-of-term decisions follow the shape of real enterprise asset-finance workflows.

Billing by the hour of use

Celery workers turn live equipment telemetry into invoices at the close of every cycle.

A price for the return

A regression model values each machine at the end of its lease, so remarketing opens with a number instead of a guess.

An analyst on call

Risk scoring, maintenance prediction and a conversational assistant run on Groq's Qwen3-32B.

The strata

How it stands
01 · FRONTENDNext.js 15, TypeScript, Tailwind v4 and TanStack Query dashboa…fig. I02 · APIDjango 5 and DRF across originations, servicing, payments and …tav. II03 · ASYNC ENGINECelery on Redis: telemetry ingestion and invoice generationc. III04 · INTELLIGENCEscikit-learn resale valuation and a Groq-hosted AI enginenota IV05 · INFRASTRUCTUREPostgreSQL 16 and Nginx, orchestrated with Docker Compose, liv…fol. VAssetStream, delineavit F.K.R.

  1. Frontend

    Next.js 15, TypeScript, Tailwind v4 and TanStack Query dashboards

  2. API

    Django 5 and DRF across originations, servicing, payments and remarketing

  3. Async engine

    Celery on Redis: telemetry ingestion and invoice generation

  4. Intelligence

    scikit-learn resale valuation and a Groq-hosted AI engine

  5. Infrastructure

    PostgreSQL 16 and Nginx, orchestrated with Docker Compose, live on Render

The judgments

The reasoning, on the record

    Chose

    Celery and Redis workers

    over billing inside the request cycle

    Telemetry arrives continuously and invoices close per cycle. A queue keeps the API responsive and lets billing jobs retry safely.

    Chose

    A scikit-learn regression

    over a deep model

    Resale value turns on a handful of tabular features. A regression explains itself to a finance team and retrains in seconds.

    Chose

    Docker Compose

    over Kubernetes

    Six containers on one host did not need an orchestrator. Raising the whole stack with a single command mattered more.

The cabinet

Captured from the live build
AssetStream, screenshot 1

AssetStream, plate I

Screen capture, 2560 × 1600

assetstream-frontend.onrender.com, MMXXVI

Next exhibit — IV

AskVid-Pro