IT
Irfan Thanveer
Multi-Agent Systems & Enterprise RAG
DISPATCH // AI CoE BENGALURU/3+ Years in Applied AI & Data Engineering/OPEN TO NEW OPPORTUNITIES
Irfan Thanveer - AI Engineer

AI ENGINEER //

MULTI-AGENT SYSTEMS

& PRODUCTION RAG

Building Enterprise-Grade Multi-Agent Systems, LangGraph Orchestration & Production LLM Platforms at Infosys BPM AI CoE.

HACKATHON WINNER2025
1st Place // 20+ Teams

Future Talent & Innovation Team Award for LangGraph Multi-Agent Insights Platform.

INNOVATIONENTERPRISE
Vectorless RAG Engine

Page-Indexed architecture with OpenAI & Claude for 100% table and figure precision.

AUTONOMOUSAZURE AI
Browser-Use Intelligence

Autonomous web agents reducing enterprise scraping and data curation overhead by 80%+.

Launch Agent PlaygroundExplore Flagship Case Studies
INTERACTIVE AGENTIC PLAYGROUND

AGENT CONSOLE & LANGGRAPH VISUALIZER

Select a scenario below or test the multi-agent graph to simulate real-time state orchestration, tool execution, and vectorless retrieval.

ORCHESTRATION TOPOLOGYLANGGRAPH

Visual state transitions for the Hackathon-winning Multi-Agent Enterprise Insights Platform:

Supervisor Planner
Intent parsing & task routing
01
Data Profiler & SQL
Schema extraction & cleaning
02
Visual Chart Engine
Distribution & trend charts
03
Executive Synthesizer
Structured decision memo
04
State: READY1st Place Award Winner
irfan@agent-cluster-01:~$
RUNTIME: LANGGRAPH|STATUS: LIVE
Quick Scenarios & Architecture Traces:
// System initialized. Click any preset above or submit a query to simulate multi-agent orchestration.
irfan-agent > LangGraph runtime online. Ready for execution.
>
ENGINEERING ARCHITECTURES & SYSTEMS

FEATURED PROJECTS

Agentic AI & LangGraph🏆 Future Talent & Innovation Award
1st Place Winner / 20+ Teams

Multi-Agent Enterprise Insights Platform

// Autonomous multi-agent system for automated data profiling, visual analytics, and contextual synthesis.

Led a team of 3 engineers to architect an autonomous enterprise insight engine powered by LangGraph. The system coordinates specialized agents for dataset schema validation, SQL querying, statistical profiling, chart generation, and executive memo synthesis.

Key Architectural Accomplishments:
  • ▪Secured 1st Place out of 20+ engineering teams at Infosys Internal AI Hackathon.
  • ▪Won the Future Talent & Innovation Team Award.
  • ▪Designed dynamic state management graphs with conditional routing and human-in-the-loop validation.
  • ▪Automated complete end-to-end data transformation to executive chart generation.
Pipeline Flow:
01. Supervisor AgentParses user query, inspects schema context, delegates tasks to specialist subagents.
02. Data TransformerExecutes automated cleaning, joins, and statistical profiling via Pandas/SQL.
03. Chart Generation AgentSynthesizes data distribution into interactive, high-contrast visual charts.
04. Executive SynthesizerGenerates natural language narrative findings with verifiable citations.
LangGraphPythonAzure AIOpenAIFastAPIPandasMatplotlib
Read Full Case Study
RAG & Enterprise LLM
100% Figure/Table Context Precision

Page-Indexed Vectorless RAG Platform

// Enterprise Q&A and process training platform using Page-Indexed retrieval to overcome vector embedding drift.

Architected a next-generation enterprise knowledge retrieval platform. Replaced traditional semantic chunk-based embeddings with structured document page indexers, OpenAI Responses API, and Claude to maintain 100% table and figure context accuracy.

Key Architectural Accomplishments:
  • ▪Eliminates retrieval degradation and hallucination on complex multi-column PDF SOPs.
  • ▪Integrated OpenAI Responses API and Anthropic Claude for structured entity extraction.
  • ▪Implemented hierarchical page metadata tagging and deterministic citation mapping.
  • ▪Supports enterprise process onboarding and technical SOP training at scale.
Pipeline Flow:
01. Document Page IndexerPerforms layout-aware structural parsing and page-level metadata indexing.
02. Intent & Page RouterIdentifies target sections and page clusters without embedding distortion.
03. Deep Extraction LLMProcesses full page contexts with Claude 3.5 / OpenAI Responses API.
04. Citation VerifierValidates response accuracy against source page coordinates.
Vectorless RAGLangGraphClaude 3.5 SonnetOpenAI Responses APIPythonFastAPI
Read Full Case Study
Agentic AI & Scraping
80%+ Reduction in Manual Extraction

Autonomous Web Intelligence Scraping Agent

// Intelligent web automation and dynamic product data extraction running on Azure AI Foundry.

Developed an autonomous browser agent combining Browser-Use, LangChain, and Azure AI Foundry. The agent navigates complex, dynamic web applications, executes anti-bot resilient scrapers, and extracts structured enterprise intelligence at scale.

Key Architectural Accomplishments:
  • ▪Automated large-scale product intelligence workflows across complex SPAs and interactive portals.
  • ▪Deployed on Azure AI Foundry for scalable enterprise execution and observability.
  • ▪Built resilient error recovery, visual grounding, and automated DOM traversal logic.
  • ▪Reduced manual data extraction overhead by over 80%.
Pipeline Flow:
01. Visual GrounderAnalyzes DOM coordinates and interactive viewport elements in real time.
02. Action PlannerFormulates sequence of clicks, form fills, and navigation events.
03. Extraction ValidatorParses returned payloads into normalized JSON schemas.
04. Azure Foundry RelayStreams structured intelligence to downstream enterprise data lakes.
Browser-UseLangChainAzure AI FoundryPythonPlaywrightDocker
Read Full Case Study
GenAI & Analytics
Streamlined Ticket Resolution

GenAI Internal Support Ticket Intelligence

// Automated ticket workflow intelligence and recurring issue root-cause analysis.

Built a generative AI solution at HfactoR to streamline support ticket resolution. Extracted unstructured ticket logs from PostgreSQL, performed automated categorization and topic clustering, and exposed operational bottlenecks via an interactive Streamlit dashboard.

Key Architectural Accomplishments:
  • ▪Automated extraction and clustering of thousands of customer support tickets.
  • ▪Built interactive Streamlit BI dashboard for engineering and operations managers.
  • ▪Engineered automated feature pipelines to flag recurring production incidents.
PythonGenAIPostgreSQLStreamlitPandasNLP
Production Internal System
Machine Learning
High Recall Loan Default Prediction

Credit Risk ML Classification Engine

// Supervised ML classification pipeline to predict loan default risks with high recall.

Developed predictive machine learning models to identify high-risk loan applicants. Implemented Random Forest, XGBoost, and AdaBoost classifiers combined with SMOTE oversampling to tackle severe class imbalance.

Key Architectural Accomplishments:
  • ▪Engineered robust feature transformation pipelines on high-dimensional tabular financial data.
  • ▪Optimized ROC-AUC and Recall metrics using ensemble tree architectures.
  • ▪Performed cross-validation and hyperparameter tuning for maximum generalization.
PythonScikit-LearnXGBoostRandom ForestPandasSMOTE
Production Internal System
Machine Learning
84.2% Valuation Accuracy

Laptop Price Predictive Valuation Model

// Supervised regression model achieving 84.2% valuation accuracy.

End-to-end regression modeling pipeline to estimate laptop market pricing based on hardware specs, brand value, form factor, and market trends.

Key Architectural Accomplishments:
  • ▪Achieved 84.2% test prediction accuracy.
  • ▪Extensive EDA, categorical encoding, and feature scaling.
  • ▪Packaged as a lightweight evaluation script and API service.
PythonScikit-LearnNumPyMatplotlibSeaborn
Production Internal System
CAREER DISPATCH // STORYSTREAM TIMELINE

PROFESSIONAL EXPERIENCE

DEC 2024 — PRESENT
Bengaluru, India

Infosys BPM

AI Engineer — AI CoE (Centre of Excellence)
AI CoE
  • ▪Leading development of next-generation enterprise Q&A and process training platforms using LangGraph, OpenAI Responses API, Claude 3.5, and Page-Indexed Retrieval Architecture (Vectorless RAG).
  • ▪Led a 3-person engineering team to build a multi-agent enterprise Insights generation system with LangGraph, winning 1st Place at the Internal AI Hackathon (20+ teams) and the Future Talent & Innovation Team Award.
  • ▪Engineered intelligent web automation agents using Browser-Use, LangChain, and Python on Azure AI Foundry for scalable enterprise web intelligence.
LangGraphVectorless RAGAzure AI FoundryBrowser-UseClaudeFastAPIPython
MAY 2024 — DEC 2024
Bengaluru, India

HfactoR

Software Engineer
GenAI & Analytics
  • ▪Built a GenAI application to streamline internal support ticket workflows by extracting data from PostgreSQL and identifying recurring operational bottlenecks.
  • ▪Delivered real-time operational insights through an interactive Streamlit dashboard.
  • ▪Performed exploratory data analysis, feature engineering, and reporting pipeline development for business stakeholders.
GenAIPostgreSQLStreamlitPythonFeature EngineeringData Analytics
JUN 2023 — NOV 2023
Bengaluru, India

iSocrates LLC

Data Engineer
Cloud Data Engineering
  • ▪Engineered Python ETL scripts to optimize high-volume data pipelines and improve execution efficiency.
  • ▪Implemented AWS Redshift, S3, and Athena for scalable cloud data warehousing and querying.
  • ▪Modernized legacy SSIS / SSMS workflows into cloud-native, maintainable Python-based pipelines.
PythonAWS RedshiftAWS S3AWS AthenaETLSQL
NOV 2020 — MAR 2022
Remote / Singapore

Project and Business Management Pte Ltd

Project Engineer (Analytics)
Operations Analytics
  • ▪Managed project planning and execution with a primary focus on operations analytics, inventory control, and process optimization.
  • ▪Developed reporting routines and visualizations using Excel, MySQL, and Python scripts.
PythonMySQLExcel AnalyticsProcess OptimizationData Modeling
TECHNICAL ARSENAL & PROFICIENCIES

SKILLS & CAPABILITIES

Specialized in productionizing AI agents, scalable FastAPI backends, and deterministic enterprise RAG pipelines.

AI & Agentic Systems10 TOOLS

AI & Agentic Systems

Agentic AILangGraphLangChainMulti-Agent OrchestrationVectorless RAGOpenAI Responses APIAnthropic ClaudeBrowser-UsePrompt EngineeringAzure AI Foundry
Production Verified
Backend & Systems10 TOOLS

Backend & Systems

Python (Async/OOP)FastAPIRESTful APIsSQLPostgreSQLMySQLMongoDBMilvus Vector DBDockerKubernetes
Production Verified
Cloud & Infrastructure6 TOOLS

Cloud & Infrastructure

Azure AI ServicesAWS RedshiftAWS S3 & AthenaGit & CI/CDETL PipelinesCloud Data Storage
Production Verified
Data Science & Analytics7 TOOLS

Data Science & Analytics

Pandas & NumPyMachine LearningRandom Forest & XGBoostNLP & Text ExtractionStreamlitFeature EngineeringStatistical Analysis
Production Verified
HONORS, CREDENTIALS & ACADEMIA

AWARDS & EDUCATION

🏆 1ST PLACE WINNER // 20+ TEAMSINFOSYS BPM AI HACKATHON

Future Talent & Innovation Award

Secured first place among 20+ competing engineering teams at Infosys BPM for designing and leading the development of the Multi-Agent Enterprise Insights Platform using LangGraph. The solution autonomously handles dataset transformation, profiling, visual chart generation, and executive memos.

TEAMS COMPETING20+ TEAMS
TEAM LEADERSHIP3 ENGINEERS
CORE RUNTIMELANGGRAPH
Explore Award-Winning Architecture
INDUSTRY CERTIFICATIONSVERIFIED

Microsoft Certified: Azure AI Fundamentals (AI-900)

Microsoft

Azure AI

Google Data Analytics Professional Certificate

Google

Google
ACADEMIC FOUNDATIONENGINEERING

Great Lakes Institute of Management

2022

Post Graduate Program in Data Science & Engineering

SRM Institute of Science and Technology

2020

B.Tech in Mechanical Engineering

ESTABLISH COMMUNICATION // DIRECT CHANNEL

LET'S BUILD SOMETHING EXTRAORDINARY

CURRENT DISPATCH LOCATION
Bengaluru, India

Available for Senior AI Engineer role

Email Addressirfaunthaanvirr@gmail.com
Direct Phone+91-95135 21591
GitHub Profile ↗LinkedIn Profile ↗

Curriculum Vitae (PDF)

Comprehensive technical resume and project breakdown.

Download PDF
MESSAGE DISPATCHDIRECT INBOX

Send an Engineering Inquiry

Copied to clipboard!