Currently Building — AI Cloud PACS · Astraios Software

Recently Completed - Global Pharma M&A — 37 Countries — Delivered 3 Months Early

I don't advise
from a deck.
I build.

Deb Pramanik — 20+ years in global IT transformation, now deploying that pattern recognition across two active ventures: an AI Cloud PACS platform for diagnostic radiology, and a wealth advisory practice for long-game investors. Senior advisory still open — selectively.

Countries led

28+

Years in IT

81%

PACS accuracy

<5min

Report time

// The field agent belief

The people who transform industries are never the ones with the best frameworks. They are the ones who have made real decisions under real constraints — and lived with the consequences.

See What I'm Building → Book an Audit

Entrepreneur · Builder · Advisor

Ex-Accenture VP · Ex-Capgemini Sr. Director · MIT

// Currently building

Two Ventures. One Mind.

Active development — not advisory. These are decisions I am living with, not frameworks I am selling.

Venture 01 · Healthcare AI

Live

Astraios Technologies

AI Cloud PACS Platform

Cloud-hosted AI Picture Archiving and Communication System for diagnostic radiology. Google Med-Gemma SLM on Azure. Purpose-built for India's diagnostic infrastructure — where radiologist capacity is the bottleneck, not imaging hardware.

Diagnostic accuracy 81% → targeting 84%+

> Report time: 12–15 min → sub 5 min
> Cost: ~25% of radiologist fee
> Investor consortium: doctors & diagnostic operators
> Pathology AI predictive modelling — in progress

Platform Partnership Enquiry →

Venture 01 · Next Module

In Development

Pathology AI

Predictive Modelling for Path Labs

India's path labs have penetrated deeper into diagnostic infrastructure than any other segment. Predictive modelling layered onto existing lab workflows — before the sample reaches the radiologist.

> Predictive pattern recognition
> Path lab workflow integration
> Early-stage diagnostic flagging
> Status: architecture & modelling phase

Updates via newsletter →

Venture 03 · Undisclosed

Next Vertical

Undisclosed

A third application of the same pattern: deep domain expertise + AI infrastructure + regulated environment. Details withheld until foundation is set.

Watch this space

Venture 02 · Financial Advisory

Sagestone — Wealth Advisory

SEBI-registered MF distribution and wealth advisory practice. NISM certified. PGDSM in progress. The same analytical rigour applied to large-scale IT programmes, applied to long-game personal wealth.

Steady Capital Newsletter → Wealth Advisory Enquiry

// Track record

Proof of Method, Not Identity

The Sandoz engagement is not a credential to be listed. It is evidence of a working method: managing complexity at scale, protecting what cannot break, moving faster than the timeline demands. The same discipline runs every venture I build.

"That is not a credential. It is a proof of method." — LinkedIn About

Anchor Engagement · Case Study

Novartis/Sandoz — 37-Country IT Separation

Post-demerger IT separation across 37 countries and 14 GxP-certified manufacturing sites. Three MDM platforms separated. ~400 OQ+PQ test scripts. Zero operational disruption. Delivered 3 months ahead of schedule.

Countries

GxP Sites

−3mo

Early delivery

Zero

Disruption

Read Full Case Study →

Senior Advisory · Available Selectively

The 30-Day AI Blueprint

Stop the "Rip and Replace" cycle. Map your legacy constraints and build the specific bridge architecture to Agentic AI — GxP-compliant and board-ready. Two slots per month.

> Week 1: Legacy debt & risk mapping
> Week 2: AI readiness assessment
> Week 3: GxP & compliance validation
> Week 4: Board-ready AI roadmap

Request the Blueprint →

Background

> Ex-Accenture Vice President

> Ex-Capgemini Senior Director

> MIT — Executive Education

> $100M+ programmes led

> Pharma, Life Sciences, MedTech

// Executive AI Workshops

Learn From Someone Currently Deploying It

Not theory. Not slides borrowed from a vendor. Two programmes built from active deployment of AI in regulated environments.

Programme 01 · Technical

3 Days

Deploying AI in Regulated Environments

From model selection to GxP validation to production deployment. Built from the Astraios build — real decisions, real infrastructure constraints, real clinical-grade requirements.

Who this is for

CTOs, Heads of Engineering, AI/ML leads in pharma, healthcare, and regulated industries.

> SLM vs LLM selection for regulated use
> GxP validation frameworks for AI
> Azure / cloud architecture for clinical AI
> Accuracy benchmarking & audit readiness
> Legacy data as AI fuel, not obstacle

Enquire → Agenda PDF

Programme 02 · Business

2 Days

AI Strategy for Business Leaders

How to lead an AI transformation without being led by your vendors. For the CXO who needs to own the room — not defer to it.

Who this is for

CEOs, CFOs, Board members, GCC heads, and senior business leaders making AI investment decisions.

> How to evaluate AI vendor claims
> Legacy debt — the ROI blocker nobody names
> Building the board-ready AI business case
> M&A AI due diligence framework
> Governance, compliance, and AI risk

Enquire → Agenda PDF

Both programmes available in-person (Bengaluru / client site) or virtual. Custom cohort formats available for organisations.

Custom Format Enquiry

// Featured whitepaper

The AI Archeology Framework: Excavating Pharma's Legacy Past to Feed Its AI Future

Large Language Models are only as smart as the data they can reach. This whitepaper maps the hidden legacy debt blocking AI ROI across pharma and life sciences — and the bridge architecture to fix it without ripping out what took 30 years to build.

Finding 01

20% of legacy debt blocks 80% of AI ROI — and most organisations cannot name which 20%.

Finding 02

GxP validation creates unique AI compliance risk that generic cloud AI guidance does not address.

Finding 03

"Rip and replace" cycles stall at year two. Bridge architecture delivers faster and with less disruption.

Finding 04

Data dark matter — unstructured, undocumented legacy data — is pharma's most underutilised AI asset.

Finding 05

Post-M&A IT separation creates a 3–18 month AI readiness window that most organisations miss.

Finding 06

SLMs tuned on domain data outperform general LLMs in clinical accuracy at a fraction of the cost.

No spam. One confirmation email. Unsubscribe any time.

// Resources

Open-Source Frameworks

Practical tools from 28 years of delivery. Download free — no login required.

📄

Legacy Audit Heatmap

Identify which 20% of legacy debt is blocking 80% of AI ROI. Built from the Sandoz post-demerger playbook.

Download PDF →

📊

AI ROI Calculator

Model the actual return on your AI investment before committing budget. Interactive Excel, pharma-first.

Download XLS →

📐

AI Architecture Guide

Bridge architecture patterns for connecting legacy pharma infrastructure to modern AI pipelines. GxP-aware.

Download PDF →

🏭

Pharma Case Study

One-page summary of the Novartis/Sandoz 37-country IT separation. Methodology, outcomes, and lessons.

Download PDF →

// Writing

Two Registers. One Perspective.

Stream 01 — Steady Capital · Substack

AI strategy, legacy debt, pharma transformation & market thinking.

Archive //

Loading…

Read on Substack →

Loading…

Read on Substack →

Stream 02 — Building in Public · LinkedIn

The PACS build journal — technical decisions, deployment lessons, and what the textbooks don't tell you.

Post series

Why We Chose Med-Gemma Over GPT-4

Clinical accuracy, cost-per-inference, and why general LLMs fall short at 84%+ threshold.

Read on LinkedIn →

Post series

When Azure India Had No Vertical Medical LLM on SaaS

Building our own Azure Container App deployment from scratch. What we learned.

Read on LinkedIn →

Post series

From 78% to 81%: The Accuracy Gap That Matters

What the jump from 81% to 84% actually requires — and why most clinical AI stops before it gets there.

Read on LinkedIn →

Full LinkedIn series //

Get the weekly dispatch — Steady Capital

AI strategy, pharma IT insights, and market thinking. Free. Plain-talk.

Subscribe Free →

// Let's talk

Start a Conversation

Advisory Enquiry

Pharma IT Transformation / 30-Day Audit

M&A IT separation, AI strategy, post-demerger programme leadership. Two slots per month.

Workshop Enquiry

Executive AI Workshops

Technical (3 days) or Business (2 days) programmes. In-person or virtual. Custom cohorts available.

Platform Partnership

Astraios / Healthcare AI

Diagnostic centres, path labs, radiology networks interested in AI Cloud PACS deployment.

Direct email

deb@sagestone.in

Location

Bengaluru, India

Available globally — remote advisory, in-person workshops

Connect

LinkedIn Substack

Current availability

Taking on 2 Legacy-to-AI Gap Audits per month. Next available slot: June 2026

I don't advisefrom a deck.I build.

Two Ventures. One Mind.

Astraios Technologies

Pathology AI

Next Vertical

Sagestone — Wealth Advisory

Proof of Method, Not Identity

Novartis/Sandoz — 37-Country IT Separation

The 30-Day AI Blueprint

Learn From Someone Currently Deploying It

Deploying AI in Regulated Environments

AI Strategy for Business Leaders

The AI Archeology Framework: Excavating Pharma's Legacy Past to Feed Its AI Future

Open-Source Frameworks

Two Registers. One Perspective.

Start a Conversation

I don't advise
from a deck.
I build.