Start a Project

Dwarka Mor, New Delhi
[email protected]

Engineers working together on a complex machine assembly

Home/Case Studies

Case Studies

Selected programmes and products from the lab — what made the problem hard, how we approached it, and what it became.

Selected Work

Problem, approach, outcome

Each study starts from the constraint that made the work hard. Where an engagement is confidential the client stays unnamed — the engineering does not.

Security cameras mounted on a building wall

Turning existing camera networks into an investigative tool

Investigation and security teams had cameras everywhere and time for almost none of the footage. ADVA, our deep vision analytics platform, brings intelligence to the camera systems agencies already run.

Challenge

Footage from large camera estates was reviewed by hand, after the fact, if at all. Replacing the cameras was not an option — the intelligence had to work with the systems already in the field.

Approach

Vision analytics built to sit on top of traditional camera systems, turning them into intelligent security networks, with a forensic review workflow for investigators.

Outcome

Our smart surveillance and video-forensics products have been rolled out to investigation and security teams across multiple agencies since 2021.

C++PythonOpenCVONNX
A processor seated on a computer motherboard

Deep hardware diagnostics for OEM production lines

OHDT, our Onboard Hardware Diagnosis Tool, tests and validates the critical components of a computer — on the factory floor and in the field.

Challenge

Computer OEMs and service teams need to know quickly and reliably whether a machine's components are sound: in factory testing, on the production line, in RMA analysis and at a customer site, across widely varied hardware.

Approach

System-level diagnostics written close to the hardware that run either inside the operating system or standalone from a bootable USB, so one test suite follows a unit from the factory to field service.

Outcome

Our hardware-diagnostic tooling is in use on production lines with electronics OEMs, and has driven the growth of our hardware engineering practice since 2024.

CC++PythonQt
Network cables plugged into a patch panel in a server rack

Domain intelligence for security investigations

Intelli brings real-time data collection, historical analysis and threat intelligence on domains together in a single investigative platform.

Challenge

Answering a simple question about a domain — who is behind it, what it connects to, what it has done before — meant stitching many separate data sources together by hand.

Approach

A platform that integrates real-time DNS collection with historical records and threat-intelligence sources, and applies analysis across them so patterns surface without manual correlation.

Outcome

Security professionals and researchers get one place to assess domain security and infrastructure, and to make informed decisions from it.

PythonFlaskDNS
A laptop, phone and contract document on a desk

AI across the full tendering cycle

Bidhelp applies AI to the whole tendering process, from reading the RFP to taking part in the tender.

Challenge

Tender documents are long, dense and time-bound. Bids are lost not on capability but on the hours it takes to find, read, qualify and respond to the right opportunities.

Approach

LLM-based analysis of RFP documents combined with a management platform that carries each opportunity from analysis through to automated tender participation.

Outcome

One platform covering the tendering cycle end to end, from RFP analysis to automated participation.

PythonNode.jsMySQLPHP CILLM
A rocket lifting off from its launch pad at night

Simulating powered vertical landing

SimX-V1 is a physics-based rocket flight simulator for developing, testing and validating guidance, navigation and control algorithms.

Challenge

Control algorithms for vertical takeoff and powered reverse landing cannot be tuned on real hardware first — every failed test is expensive and unrecoverable.

Approach

A physics-based flight simulator that models the vehicle through takeoff and powered descent, so GNC algorithms can be developed, stressed and validated in software before they ever meet hardware.

Outcome

A working environment for developing and validating autonomous vertical takeoff and powered reverse-landing control.

CC++Python
A software engineer working at monitors full of code in a dim room

An enterprise software suite for growing businesses

CRM, HRMS, POS and Garage108 — business software products built by the lab for private-sector enterprises.

Challenge

Growing businesses need dependable systems for customers, people and sales, without the cost and complexity of software designed for large corporations.

Approach

A family of focused products — CRM, HRMS, POS and Garage108 — built to the same engineering standards as our work for government and defence.

Outcome

Released in 2022 and in use across hundreds of business deployments, taking the lab into private-sector enterprise software.

Streaks of blue light across a dark background

Your Programme

Have a problem like one of these?

Tell us what you are building and what is making it hard. We will tell you which of these approaches it resembles — and where it differs.