Booting the keyboard

Available for AI Engineering roles · Interactive keyboard nav

Kunal Gunjan Talukdar

AI Engineer

Building AI Systems, LLM Applications, Fine-tuning, Agentic Workflows.

I build production-ready AI systems that combine machine learning, large language models, and software engineering.

Tip · press the keys on the 3D keyboard

About

Data & AI professional shipping AI-powered systems end-to-end.

Data / Product Analyst with 6+ years in fintech — regulatory, lending and payments — at Slice (formerly North East Small Finance Bank). I build and ship AI-powered systems end-to-end: a RAG-based financial data query system, Slack-integrated AI agents, and internal MCP tooling, alongside applied LLM post-training work (SFT, LoRA/PEFT, evaluation). Recognized as Slice’s AI Champion for Q3 FY25.

AI & Agentic Systems

LangChain, LangGraph, RAG pipelines, vector databases, MCP tooling and Slack-integrated agents.

Applied ML & Post-Training

SFT, LoRA / PEFT with Thinking Machines’ Tinker; evaluation and benchmarking (LiveSQLBench).

Data, Analytics & Product

Python, SQL, Spark and Airflow on AWS — turning fintech data into product decisions.

LLMsFine-tuningRAGLangChainLangGraphMCPAgentic AIPythonSQLApache SparkAWSAirflow

Experience

Six+ years across fintech, analytics and applied AI.

From regulatory analytics at a scheduled bank to shipping Slack-integrated AI agents and MCP tooling at Slice.

Apr 2026 — Present

Analyst, Product – Consumer Lending

Slice

  • Own analytics, reconciliation, and RBI regulatory reporting for credit card and other digital payment products within the consumer lending line of business.
  • Working at the intersection of product and data to bring the team’s AI-agent and automation tooling into the consumer lending workflow.

May 2025 — Apr 2026

Analyst, Data Analytics

Slice (formerly North East Small Finance Bank)

  • Built AI agents integrated with Slack for data analysis and workflow automation, using LangChain, LangGraph, and a vector database for retrieval.
  • Automated data-analytics workflows — including JIRA ticket creation — by integrating AI agents via n8n.
  • Designed and built a RAG-based financial data query system (n8n, PostgreSQL, JavaScript Code Nodes for Bayesian disambiguation, OpenAI embeddings, Slack interactive messages) that outperformed an existing n8n-based competitor system.
  • Developed an MCP server for retrieving files from a shared drive, and drove team-wide AI adoption by rolling out Cursor and MCP tooling.
  • Delivered analytics across the LMS, LOS, loan utilization, and core banking systems.
  • Recognized as Slice’s AI Champion for Q3 FY25 (Jul–Sep 2025) for AI-driven workflow automation.

Sep 2019 — Apr 2025

Sr. Digital Marketing Executive (prev. Digital Marketing Executive)

North East Small Finance Bank

  • Delivered Regulatory (RBI) Analytics and Marketing Analytics alongside content and digital-channel responsibilities; promoted to Sr. Digital Marketing Executive in Sep 2023.
  • Piloted early Slack-based workflow automation in 2024, laying the groundwork for the AI agent build-out later carried forward at Slice.
  • Managed social media content, SEO, and brand presence across digital channels for a scheduled bank.

Jun 2018 — Aug 2019

Digital Marketing Executive

Innfinity · Ezen Software and Technology Pvt. Ltd.

  • Managed digital marketing and content execution for client accounts.

Jul 2017 — Jun 2018

Digital Marketing Executive

Zantrik Technologies Pvt. Ltd.

  • Started career in digital marketing, building foundational content and campaign-execution skills.

Education

B.Tech, Computer Science & Engineering

GIMT, Azara · 2013 — 2017

Projects

Selected work.

AI systems shipped across fintech analytics, agentic workflows and applied LLM research.

RAG · n8n · PostgreSQL

RAG-Based Financial Data Query System

Conversational query system for internal financial data — pure n8n architecture with PostgreSQL storage, JavaScript Code Nodes for Bayesian disambiguation, OpenAI embeddings for semantic matching, and Slack interactive messages. Outperformed the existing competitor n8n system while keeping zero maintenance overhead.

LangChain · LangGraph · MCP

AI Agent Workflow Automation

Slack-integrated AI agents (LangChain, LangGraph, vector DB) plus n8n automations — including JIRA ticket creation — and an MCP server for shared-drive file retrieval. Drove team-wide adoption of Cursor and MCP tooling. Recognized as Slice’s AI Champion for Q3 FY25.

SFT · LoRA / PEFT · Evaluation

LLM Post-Training & Fine-Tuning

Hands-on post-training work with Thinking Machines’ Tinker — SFT and LoRA fine-tuning on Qwen3.5-4B and Kimi K2.6. Building training pipelines across forward/backward passes, optimizer updates and inference; benchmarking against LiveSQLBench.

MCP · Cursor · Internal tools

MCP Tooling & Adoption

Custom MCP server for retrieving files from a shared drive, plus adoption playbooks that brought Cursor and MCP into the team’s daily workflow — unblocking analysts and data engineers.

Coming soon

Future AI Projects

A pipeline of upcoming AI systems — next up: fine-tuning Inkling for text-to-SQL generation. Spanning research, agentic products and open-source contributions.

Skills

The stack behind the work.

A pragmatic toolkit spanning data, applied AI, product analytics and fintech domain.

Data & Programming

PythonSQLApache Spark

Cloud & Data Infra

AWSEMRS3Apache Airflow

BI & Visualization

Apache SupersetTableau

AI & Agentic Systems

LangChainLangGraphRAGVector DatabasesPrompt EngineeringModel Context Protocol (MCP)

Applied ML / Post-Training

Supervised Fine-Tuning (SFT)LoRA / PEFTThinking Machines’ TinkerModel EvaluationLiveSQLBench

Automation & Tooling

n8nCursorGitJIRASlack API

Product & Domain

Product AnalyticsConsumer LendingCredit Card & Digital PaymentsRBI Regulatory ReportingLMS / LOSCore Banking Systems

Certifications

Licenses & certifications.

Formal training in modern AI, agentic systems and machine learning fundamentals.

Model Context Protocol: Advanced Topics

Anthropic · Oct 2025

MCPModel Context Protocol

Introduction to Model Context Protocol

Anthropic · 2025

MCPModel Context Protocol

MCP: Build Rich-Context AI Apps with Anthropic

DeepLearning.AI · Oct 2025

Model Context ProtocolMCP

Foundation: Introduction to LangGraph

LangChain · Sep 2025

LangGraphAI Agents

Retrieval Augmented Generation (RAG)

DeepLearning.AI · Aug 2025

RAGGenerative AI

Machine Learning Specialization

Stanford University · Coursera · 2025

Machine Learning

Unsupervised Learning, Recommenders, Reinforcement Learning

Stanford University · Coursera · 2025

Anomaly DetectionReinforcement Learning

Advanced Learning Algorithms

Stanford University · Coursera · Mar 2025

TensorFlowArtificial Neural Networks

Supervised Machine Learning: Regression and Classification

Stanford University · Coursera · Jan 2025

Linear RegressionLogistic Regression

Contact

Let’s build something intelligent.

Open to Data / AI roles, research collaborations and consulting.

Bangalore, India