Back to site
Selected work

A new venture. Not a new team.

Doing AI is new — but its founder, Nithish Singh, has shipped production systems for years across manufacturing, education, AI and infrastructure. Every project below was designed and built by him, end to end.

Real, delivered projects — each built end to end by our founder, Nithish Singh (previously as Adidev Industries and consulting engagements). Some clients are unnamed under confidentiality; the work is real, and referenceable on request.

Work with us
ManufacturingAutomotiveQuality control
Client work

Stage-gate QC for an LCV assembly line

A major Indian commercial-vehicle manufacturer running light-commercial-vehicle assembly on paper quality checks — faults and missing parts slipping downstream, with no traceability.

The problem

  • Paper checklists at every assembly stage
  • Faults found late, at final inspection
  • Missing parts reaching downstream stations
  • No record of where a defect entered

What we built

  • A QR code at every station and part — scan to open that stage’s checks
  • Part verification, fault logging and missing-part flags, filled on the spot
  • A unit can’t advance until its stage checks pass
  • Full digital traceability for every vehicle, every stage
QR
scan-to-check at every stage
Per-unit
full traceability
Live
fault & missing-part flags
Client name withheld under confidentiality. Built end to end by our founder, Nithish Singh.
EducationPlatformProduction
Client work

UOS — engineering a University Operating System

A production “University Operating System” running admissions and operations for 11+ Indian universities. Brought in to ship core systems straight into live production.

The problem

  • Operations run manually and ad-hoc across many universities
  • No warehouse-backed reporting — ops ran raw SQL by hand
  • Document review and fee handling had to scale per-university

What we built

  • A tiered document-verification pilot (L1 → L2 → L3), live in prod
  • A production fee-ledger overhaul — multi-currency, partial payments
  • A data-warehouse + MIS layer — materialized views, 63-column master report
  • Full-stack university onboarding — WhatsApp + email, event-driven
127
production commits shipped
5
major workstreams, live
11+
universities on the platform
Consulting engagement, shipped by our founder, Nithish Singh. Figures from engagement reports.
Our productAI sales agentVoice + chat
Our product

LEOS — an AI sales agent for inbound leads

Inbound leads from ads and websites go cold in minutes. LEOS contacts every one within 60 seconds — across WhatsApp, SMS, AI voice and email — qualifies them in real conversation, and books the consult.

The problem

  • Leads go cold before anyone replies
  • Manual follow-up scattered across channels
  • No qualification until a human gets to it

What we built

  • Contacts every lead within 60 seconds of arrival
  • WhatsApp + SMS + AI voice + email — one agent
  • Qualifies through real conversation, then books or enrolls
  • Telecom compliance built in (India DLT/TRAI; US-ready)
<60s
to first contact
4
channels, one agent
Compliant
DLT/TRAI by design
Our own product. Built end to end by our founder, Nithish Singh.
Our productEdtechAI
Our product

Zelium — an AI-native operations platform for higher ed

Institutions stitch together five-plus tools that don’t talk to each other. Zelium is one platform for the whole student lifecycle, with AI woven through.

The problem

  • CRM, LMS, payments and documents all separate
  • No single source of truth per student
  • Manual document checks and counsellor grunt-work

What we built

  • One multi-tenant platform with strict per-tenant data isolation
  • Document AI that reads and verifies admissions paperwork
  • A Claude-backed counsellor co-pilot
  • WhatsApp-native — built for how Indian institutions actually work
1
platform, full lifecycle
Doc AI
reads admissions paperwork
Co-pilot
Claude-backed
Our own product. Built end to end by our founder, Nithish Singh.
Our productAIAgents
Our product

SolveX — an AI operating system for product development

A team of specialized AI agents — each with strict roles, enforced contracts and verification loops — that takes an idea from concept to a deployed product.

The problem

  • AI coding tools produce demos, not production software
  • No roles, no contracts, no verification — just raw output
  • Hard to trust AI-built systems end to end

What we built

  • Specialized AI agents with enforced roles and contracts
  • Verification loops so output is checked, not just generated
  • Concept to deployed product, end to end
Agents
specialized, role-bound
Contracts
enforced between them
Verified
output checked in loops
Our own product. Built end to end by our founder, Nithish Singh.
Our productSales intelligence
Our product

BillGill — call analytics & sales intelligence

Sales teams make thousands of calls with no idea which reps, scripts or leads actually convert. BillGill turns raw call data into decisions.

The problem

  • Call data siloed across separate dialers
  • No view of rep productivity or lead quality
  • Decisions made on gut, not data

What we built

  • Aggregates call data from multiple dialers into one view
  • Rep productivity, lead management and sales-performance insights
  • Dashboards that surface what’s actually working
2+
dialers unified
Rep + lead
performance insights
Live
dashboards
Our own product. Built end to end by our founder, Nithish Singh.
ConsultingInfrastructureCost audit
Client work

Production infrastructure & cost audit

Two live education platforms on AWS were quietly drifting into cost and risk. A read-only audit turned an invisible problem into a prioritized fix list — with zero downtime.

The problem

  • Servers over-provisioned and under-watched
  • One database logging tens of GB of errors, disks filling
  • No clear picture of what was costing what

What we built

  • Read-only diagnostics across live production servers
  • Verified AWS pricing and right-sizing recommendations
  • Found log floods, disk bloat and config risks
  • A prioritized cost + health fix list — no changes, no downtime
2
production platforms audited
0
downtime — read-only
Cost↓
right-sizing + cleanup
Consulting engagement. Built end to end by our founder, Nithish Singh.

See what we'd build for you.

Get an instant, tailored plan for your own workflow — or book a free 30-minute build session.

Get an instant AI plan Book a build session