Available for new engagements·2026

Toshi Consulting Services

AI-native platforms for security & governance.

From SentinalAI to Escalion to URBI — production-grade AI products across LLM security, cloud identity, governance, and adaptive assessment.

6
shipped platforms
170+
active users
100%
production-grade

What we do

Three principles, one production playbook.

Toshi Consulting Services designs and ships AI platforms across LLM security, cloud identity, governance, and assessment — architecture, agents, frontend, infra, all under one roof.

01

Security-first

From SAST/DAST pipelines to hardened deployment — security isn't bolted on, it's the product itself.

SAST · DAST · Hardened deploys
02

AI-native

LLM agents that actually ship: SQL copilots, code reviewers, adaptive assessments — grounded, evaluated, observable.

LLM agents · RAG · Evals
03

Production-grade

Type-safe, monitored, and live. Real users, real subdomains, real uptime — not a demo gathering dust on a localhost.

Live SaaS · Monitored · Type-safe

Selected Work

Six shipped platforms

Production builds across AI security, cloud identity, and applied AI — each card spins in 3D. Demos are available on request.

drag to spin

Urban governance copilot · Govt. of Haryana demo

URBI Command Center

A bilingual (English/Hindi) AI assistant for municipal commissioners. Ask any governance question — URBI generates a safe SQL query, runs it on live ULB data, and renders an interactive dashboard. Built around Llama 3.3 on Groq.

Next.jsLlama 3.3GroqPostgreSQLText-to-SQL
drag to spin

Secure code review platform

Sentinel SCR

Multi-tenant SaaS for continuous secure code review. Runs gitleaks + semgrep across customer repos, triages findings by severity, tracks suppressions, and maps each finding to CWE classifications.

Next.jsSemgrepGitleaksCWEMulti-tenant
drag to spin

ExamPrep platform · Live SaaS with revenue

Adaptive Assessment AI

AI-powered adaptive assessment platform for Indian K-12. Schools and independent students get personalized question banks, analytics, and live session monitoring. Currently serving 170+ users across 6 schools with active paid revenue.

Next.jsFastAPIPostgreSQLAdaptive MLStripe
drag to spin

AI-Native Security OS · LLM & AI supply-chain defense

SentinalAI

Real-time detection and blocking of AI threats — prompt injection, jailbreaks, training-data extraction, agent hijacking — mapped to the MITRE ATLAS taxonomy. Ships with an Auto-Response Engine and a policy engine that sits in front of LLMs, agents, and API gateways. Reduces AI vulnerabilities by 78%.

LLM SecurityMITRE ATLASAuto-ResponsePolicy EngineAI Gateway
drag to spin

Exploit-path intelligence · AWS IAM

Escalion

Identity-native attack-path analysis for AWS. Builds a directed identity graph (CAN_ASSUME, TRUSTS, HAS_POLICY, cross-account) from read-only IAM snapshots and enumerates real privilege-escalation chains — the paths CSPM/CIEM scanners miss. Deterministic, with SHA-256 content-addressed snapshots.

AWS IAMGraph AnalysisCross-AccountCloudTrailIdentity Security
drag to spin

Internal HR platform · Private deployment

HRMS Portal

Full-stack HR management portal — employee directory, attendance and leave tracking, payroll runs, and role-based admin dashboards. Built for an internal customer; not publicly accessible. Demo available on request.

Next.jsPostgreSQLRole-based AccessPayrollAttendance

The Stack

What we build with

16 tools we ship with — laid out as a honeycomb of golf-ball tiles. Hover to highlight, drag empty space to orbit.

hover · drag to orbit

Contact

Let's build something together

Have a security, AI, or governance product in mind? We'd love to hear what you're working on.