Selected work

Three engagements.
Three categories of compound.

Three engagements at different shapes of compounding — a Nasscom-recognised deeptech SaaS we own and are building (InfraOS), a connected-product line we engineered for an Indian wellness FMCG group (named on market launch), and a Vision AI pipeline we delivered into production for Reliance Industries' METL division.

  1. 01

    SFHL own SaaS · Nasscom-recognised deeptech · in active development

    InfraOS — digitising the land lifecycle

    InfraOS is SFHL's own deeptech SaaS — a GIS-based platform for the land-parcel lifecycle in Indian land-management contexts. Multi-stakeholder workflow, parcel-level audit trail, regulatory-compliance architecture. Selected by Nasscom for the Emerge 50 deeptech recognition track. Currently in active development.

    Tech stack

    • GIS mapping
    • Multi-stakeholder workflow
    • Compliance-ready data model
    • Python · FastAPI
    • PostgreSQL · TimescaleDB
    • Docker

    Outcome

    Nasscom Emerge 50 selection · platform under active development

    Read the case study
  2. 02

    Indian wellness FMCG group · two-product engagement · client name held until next launch

    Smart connected product engagements

    Two parallel product programmes for a leading Indian wellness FMCG group. Programme #1: an affordable category-killer connected product, designed to undercut on price while matching feature parity — currently in market. Programme #2: a smart plug-in product, currently in active development. Same client, same engineering team across both — full-stack delivery from concept to manufacturing readiness.

    Tech stack

    • Industrial design
    • Hardware engineering
    • Embedded systems
    • Mobile + cloud

    Outcome

    Programme #1 in market · Programme #2 engineered to manufacturing readiness

    Read the case study
  3. 03

    Reliance Industries Limited (RIL · METL)

    Vision AI for Reliance construction earthwork

    Drone + Vision AI pipeline for construction-site analytics at Reliance's METL operations — validating progress against engineering plans, replacing manual survey workflows with automated reporting. Paid engagement, in production.

    Tech stack

    • Vision AI
    • Drone integration
    • ML pipelines
    • Python · cloud

    Outcome

    Vision AI pipeline in production at Reliance

    Read the case study

Disclosure protocol

What we say publicly. What we say under NDA.

Public surface: client names + project category + technology stack + outcome category. Everything you'd need to evaluate fit. Under NDA, in a first or second meeting: specific deployment metrics, financial outcomes, named project codenames, reference contacts at the named clients.

We name where we can — most clients are happy to be referenced because the work landed. We anonymise where a client has specifically asked us to (typically because their procurement team prefers it). The default is to name; the exception is to anonymise. The first conversation is 30 minutes; reference setup happens after that.

Bring us a problem from the same shape as one of these.

Or a problem from a shape we haven't built yet. Both are interesting. The first conversation is on us.