---
title: "ashis deb - Sr.Data Scientist in Bengaluru | Tal"
description: "Meet ashis deb, Agentic AI Architect | Senior AI Engineer | AI Security | Multi-Agent Systems · RAG · MCP · LLMOps | IBM based in Bengaluru. AI stack includes cursor, hugging face, and langchain."
image: "https://tal.club/meet/ashis-deb/share-image"
---

# ashis deb on Tal

## Headline

Agentic AI Architect \| Senior AI Engineer \| AI Security \| Multi\-Agent Systems · RAG · MCP · LLMOps \| IBM

## About

I design and build enterprise Agentic AI platforms — multi\-agent systems that combine LLM reasoning, RAG, tool calling, MCP, evaluation, and production\-grade AI security\. At IBM, I lead the architecture and delivery of Agentic AI solutions across security automation and enterprise workflows, working with a team of engineers to take systems from architecture through production deployment\. What I’ve built Enterprise Agentic AI Platform: Built and deployed MCP servers exposing enterprise APIs as governed tools, integrated through ContextForge with watsonx Orchestrate\. Designed reusable multi\-agent workflows using RAG and tool/function calling across fact validation, motor insurance claims, and banking payment exceptions\. AI Security Platform: Architected an end\-to\-end multi\-agent platform spanning Architecture Review, Threat Modeling, AppSec/Code Review, and Privacy Impact Assessment\. Architecture Analysis: Converts PRDs/LLDs into context graphs, generates attack paths, and performs security analysis using hybrid deterministic \+ LLM reasoning — resolving \~30–40% of architecture gaps without LLM calls\. Threat Modeling: Maps architecture risks to STRIDE and MITRE ATT&CK, combining knowledge\-driven threat reasoning with product\-specific attack scenarios\. AppSec: Correlates application and code context with security knowledge to generate actionable findings while using persistent memory to reduce repeated noise and false positives by \~40%\. AI Security & Guardrails: Designed layered controls for prompt injection, jailbreaks, tool\-use boundaries, policy enforcement, schema validation, auditability, and validator/judge\-agent based output governance\. SOC Automation — IBM ATOM: Built specialist agents for IOC extraction, IP/domain reputation, hash validation, investigation, and summarisation\. Added LLM evaluation for faithfulness/relevance and OpenTelemetry/Grafana\-based observability\. Conversational Analytics — GenBI: Built an NL\-to\-SQL assistant using CodeLlama, Granite, and Milvus, enabling analysts to query enterprise data using natural language\. Technology LangGraph · LangChain · MCP · ContextForge · RAG · watsonx Orchestrate · GPT\-4o · Granite · Mistral · Milvus · FAISS · FastAPI · Docker · OpenShift · Kubernetes · OpenTelemetry · Grafana · MITRE ATT&CK · STRIDE · OWASP 12\+ years across AI/ML, NLP, Generative AI, and enterprise AI engineering, with additional specialization in AI security architecture\. Interested in Senior AI Engineer, Agentic AI Architect, AI Solution Architect, and AI Security Architect opportunities\.

## Location

Bengaluru

## AI stack

cursor

hugging face

langchain

langgraph

openai

ChatGPT

## Work experience

Sr\.Data Scientist at IBM, May 2024 \- Present · 2 yrs 5 mos, Bangalore Urban, Karnataka, India, At IBM, I design and build enterprise Agentic AI systems using LLMs, multi\-agent orchestration, RAG, and tool calling to automate complex security and business workflows\. My work spans AI security architecture, threat modeling using STRIDE and MITRE ATT&CK, LLM guardrails, and SOC automation, including specialist agents for IOC extraction, IP/domain validation, threat enrichment, and investigation summarization\. I also architect reusable agentic platforms using LangGraph, LangChain, MCP, watsonx Orchestrate, and OpenShift, extending the same architecture beyond cybersecurity into enterprise workflow automation\.

Senior Data Scientist at Cognizant, Feb 2022 \- Jun 2024 · 2 yrs 5 mos, Delivered Vertex AI / Gemini / Mistral solutions for an AI marketplace \(Google client\), including a RAG assistant that improved discovery of internal tools and reduced manual search effort\.

Sr\.associate at Cognizant, Feb 2022 \- May 2024 · 2 yrs 4 mos, Bangalore, Sr\. Associate — Data Scientist Cognizant · Full\-time \| Feb 2022 – May 2024 \| Bengaluru \- Delivered Vertex AI / Gemini / Mistral solutions for an AI marketplace \(Google client\) — RAG assistant that improved internal tool discovery and reduced manual search effort \- Built AI360 Insights Hub for media sentiment analysis and CV\-to\-job matching \- Developed BERT\-based SLA root\-cause classifier — \~80% F1 score \- Automated HR and payroll workflows using Flask, Snowflake, Azure, CI/CD

Senior Data Analyst at Accenture, Jun 2019 \- Feb 2022 · 2 yrs 9 mos, Built NLP and clustering pipelines for survey data modelling using TF\-IDF, K\-means and LSTM/GloVe, improving call\-centre insight generation\.

Data Scientist at Accenture, May 2019 \- Feb 2022 · 2 yrs 10 mos, bangalore

Machine Learning Engineer at SPi Global, Oct 2018 \- May 2019 · 8 mos, Chennai Area, India, Built Doc2Vec\-based research\-paper classification workflows to improve automated categorisation and document discovery\.

Data Scientist at PayZello, Jan 2017 \- Oct 2018 · 1 yr 10 mos, Hyderabad Area, India, My work here involves structure user data ,organise them according to the requirement of the project \. For extracting raw data we use mongo db query , python\(pandas,numpy\) is vastly used for data munging manipulation in data frames , dictionaries and lists objects \. We vastly use Elastic Search for storing analysed data , for fast query and analysis\. I was involved in developing dashboards using R\-shiny and plotly visualisation \. I was actively involve in developing web based reporting system in Flask and mongo db \. Recently involve in developing chatbot for our organisation , in which applying NLP and machine learning models like naive Bayes , random Forest , logistic regression , deep learning models like word2vec and doc2vec models \.

Data Scientist at Predictive research, Nov 2013 \- Oct 2016 · 2 y 11 mo, Bangalore

## Education

Btech in Information Technology\., Information Technology, GIET University Gunupur

Bachelor of Technology, Information Technology, Gandhi Institute of Management & Technology

[View the original profile](https://tal.club/meet/ashis-deb)

```json
{
  "@context": "https://schema.org",
  "@graph": [
    {
      "@type": "Person",
      "@id": "https://tal.club/meet/ashis-deb#person",
      "name": "ashis deb",
      "url": "https://tal.club/meet/ashis-deb",
      "image": "https://tal.club/meet/ashis-deb/profile-photo.jpg",
      "description": "Meet ashis deb, Agentic AI Architect | Senior AI Engineer | AI Security | Multi-Agent Systems · RAG · MCP · LLMOps | IBM based in Bengaluru. AI stack includes cursor, hugging face, and langchain.",
      "jobTitle": "Sr.Data Scientist",
      "worksFor": {
        "@type": "Organization",
        "name": "IBM"
      },
      "hasOccupation": {
        "@type": "Occupation",
        "name": "Sr.Data Scientist"
      },
      "homeLocation": {
        "@type": "Place",
        "name": "Bengaluru"
      },
      "alumniOf": [
        {
          "@type": "EducationalOrganization",
          "name": "GIET University Gunupur"
        },
        {
          "@type": "EducationalOrganization",
          "name": "Gandhi Institute of Management & Technology"
        }
      ],
      "knowsAbout": [
        "cursor",
        "hugging face",
        "langchain",
        "langgraph",
        "openai",
        "ChatGPT"
      ],
      "mainEntityOfPage": {
        "@id": "https://tal.club/meet/ashis-deb#webpage"
      }
    },
    {
      "@type": "ProfilePage",
      "@id": "https://tal.club/meet/ashis-deb#webpage",
      "url": "https://tal.club/meet/ashis-deb",
      "name": "ashis deb on Tal",
      "description": "Meet ashis deb, Agentic AI Architect | Senior AI Engineer | AI Security | Multi-Agent Systems · RAG · MCP · LLMOps | IBM based in Bengaluru. AI stack includes cursor, hugging face, and langchain.",
      "dateModified": "2026-09-11T12:00:21.780Z",
      "isPartOf": {
        "@id": "https://tal.club/#website"
      },
      "mainEntity": {
        "@id": "https://tal.club/meet/ashis-deb#person"
      },
      "primaryImageOfPage": {
        "@type": "ImageObject",
        "url": "https://tal.club/meet/ashis-deb/profile-photo.jpg",
        "contentUrl": "https://tal.club/meet/ashis-deb/profile-photo.jpg",
        "caption": "ashis deb profile photo on Tal"
      },
      "inLanguage": "en-IN"
    }
  ]
}
```