❯ ./advise --region 'kolkata,eastern-india,india' ▌
/consultingAI Consultant in Kolkata: Agentic AI, GenAI & AI Security
// Senior, hands-on advisory for leaders who need AI that actually works in production: secure, accountable, and under your control. From a Kolkata practitioner who builds these systems every day.
20+
years in technology
5M+
monthly users on a platform I led
₹6.5 Cr
annual tech budget owned
22
engineers in the org I led
Why Most AI Initiatives Stall, and How to Avoid It
Most AI projects stall between the demo and production. The model usually isn't the problem. The architecture, data, security and ownership are. I spent fourteen years with Hoichoi / SVF, ending as VP of Technology, where I led the move from third-party SaaS to a homegrown streaming platform serving 5M+ monthly users. Today, as Head of Technology at Humaineeti, I build infrastructure that lets AI agents work safely alongside people. My advice comes from that work.
AI Consulting Services
Agentic AI Architecture & Readiness
For: CTOs, engineering leaders, product teams
- ❯Pick the workflows where AI agents will pay off, and rule out the ones where they won't
- ❯Design orchestration, state management, tool access (MCP) and human-in-the-loop checkpoints
- ❯Review architecture and production readiness before you scale
GenAI Use-Case & ROI Assessment
For: CEOs, CIOs, business-unit heads
- ❯Shortlist and rank use cases by value, feasibility and risk
- ❯Build-vs-buy decisions, model selection and total cost of ownership
- ❯A pilot plan with clear success metrics and evaluation
Agentic SOC & AI Security
For: CISOs, SOC and security engineering teams
- ❯AI-assisted alert triage, investigation and response, with human approval where it counts
- ❯SIEM and firewall integration via open-source MCP servers (Wazuh, pfSense)
- ❯Threat-model your AI systems: prompt injection, data leakage, supply-chain and MCP risks
AI Governance, Privacy & DPDP Readiness
For: Boards, risk & compliance, data-protection officers
- ❯An AI usage policy, guardrails and approval workflows
- ❯Privacy-by-design data flows, informed by India's DPDP Act
- ❯Audit trails and accountability for automated decisions
Sovereign, Open-Source-First AI Infrastructure
For: Enterprises, BFSI, public sector, regulated industries
- ❯Choose between open-weight and hosted models, and between on-prem and cloud
- ❯Keep data and models under your control and avoid vendor lock-in
- ❯Build on hardened GNU/Linux and open-source foundations
Platform, Cloud & Digital Transformation
For: Media & OTT, consumer platforms, scale-ups
- ❯Move off third-party SaaS onto a platform you own
- ❯Scale, reliability and cloud-cost trade-offs
- ❯Design and lead engineering organisations through transformation
How an Engagement Works
01
Intro call
A short conversation about your goals, constraints and timelines, to see whether there's a fit.
02
Discovery
Review your current systems, data, team and risks. Interview stakeholders where useful.
03
Recommendations
A written assessment and prioritised roadmap: what to build, buy, pilot, or leave alone.
04
Hands-on support
Optional architecture reviews, pilot oversight and team coaching while you execute.
AI Consulting Across Kolkata, Eastern India & India
Kolkata & West Bengal
In-person working sessions across Salt Lake Sector V, New Town / Rajarhat, Park Street and BBD Bagh, and in Durgapur, Siliguri and Kharagpur.
Eastern & North-Eastern India
For organisations in Bhubaneswar, Guwahati, Patna, Ranchi and Jamshedpur: on-site visits where needed, remote otherwise.
Across India & remote
Remote-first engagements for teams in Bengaluru, Mumbai, Delhi NCR, Hyderabad, Pune and Chennai, with on-site workshops at key milestones.
Credibility You Can Check
- ❯I've received the Accelerator X Awards 2026 (Technology), CIO of the Year and Cyber Sentinel, among 7 industry awards
- ❯I maintain open-source Wazuh and pfSense MCP servers and Agentic SOC tooling (projects)
- ❯I've been an invited speaker at CIO500, the Bengal Chamber, IAMAI and Express Computer (speaking)
- ❯I began my career at Tata Consultancy Services, and I've worked with GNU/Linux and open source since 2003
AI Consulting in Kolkata: FAQs
❯What kind of AI consulting does Aloke Majumder offer?
I offer selective, senior advisory on Agentic AI architecture, GenAI use cases and ROI, the Agentic SOC and AI security, AI governance and DPDP-aware privacy, and sovereign, open-source-first AI infrastructure. I also advise on platform and cloud engineering and digital transformation.
❯Do you work with organisations outside Kolkata?
Yes. I'm based in Kolkata and work in person across West Bengal and Eastern India, including Bhubaneswar, Guwahati, Patna, Ranchi and Jamshedpur. Engagements elsewhere in India or abroad run remotely, with on-site sessions when they help.
❯We're a mid-sized company just starting with AI. Is this a fit?
Often, yes. A short use-case and readiness assessment is usually the right first step. It helps you avoid expensive pilots that never reach production and focus on the few use cases that will.
❯Can you help secure the AI systems we've already deployed?
Yes. Threat-modelling AI and agentic systems (prompt injection, data leakage, excessive tool permissions and supply-chain risk) is a core focus, alongside the Agentic SOC work I publish as open source.
❯Are you vendor-neutral?
Yes. Recommendations are open-source-first and vendor-neutral, and are meant to keep your organisation in control of its data, models and roadmap.
❯How do engagements start?
Email hello@alokemajumder.com or send a WhatsApp message to +91 95 9896 9896 with your organisation, context and what you'd like help with. We start with a short intro call to confirm fit and scope. I take on a small number of engagements alongside my role as Head of Technology at Humaineeti.
❯What is an Agentic SOC, and is it ready for Indian enterprises?
An Agentic SOC uses AI agents to triage alerts, gather context and draft responses, while analysts approve any high-risk action. It's ready for well-scoped work today, like alert enrichment, triage and investigation summaries, as long as you have good telemetry (for example a SIEM such as Wazuh), clear playbooks and human approval gates. I help teams start there and expand safely.
❯Is MCP (Model Context Protocol) secure enough for enterprise use?
It can be, but only with deliberate controls. MCP connects AI assistants to real tools and data, so the risks are over-broad permissions, prompt injection through tool output, and untrusted third-party servers. I recommend least-privilege tool scopes, approved server allow-lists, logging every tool call and human approval for destructive actions. I build and maintain MCP servers myself, so this advice comes from practice.
❯Does India's DPDP Act apply to AI systems and AI agents?
Yes, whenever an AI system processes digital personal data of people in India, whether in prompts, retrieval, fine-tuning or agent actions. That brings in consent, purpose limitation, data minimisation and security safeguards. I help with the technical side, such as data flows, redaction, access control and audit trails, and I work alongside your legal counsel rather than replacing them.
❯Should we hire an AI consultant or an AI development company?
Often both, but in that order. A development company builds what you specify. An independent consultant helps you decide what's worth building, how to secure and govern it, and how to judge vendors, without a stake in selling you more development hours. I can also review proposals from AI vendors and agencies on your behalf.
Talk to an AI Consultant in Kolkata
Tell me a little about your organisation, where it's based, and what you're trying to achieve with AI. I'll reply to set up a short intro call. Looking to upskill your team instead? See AI training, or book a keynote.