Akash Deep

AI Researcher

Bhagalpur, Bihar, India3 yrs 4 mos experience

Most Likely To SwitchAI ML Practitioner

Key Highlights

Engineered multi-agent LLM systems for UPI operations.
Developed enterprise-grade AI tools ensuring data security compliance.
Built fraud detection models catching 95% of fraudulent transactions.

Stackforce AI infers this person is a Fintech AI Engineer specializing in Generative AI and NLP solutions.

Contact

Skills

Core Skills

Large Language Models (llm)Artificial Intelligence (ai)Data AnalysisMachine LearningNatural Language Processing (nlp)Web DevelopmentDjangoData ScienceData EngineeringBig Data

Other Skills

API integrationAgile MethodologiesAnalytical SkillsApache SparkBitbucketBusiness AnalysisC (Programming Language)C++Cascading Style Sheets (CSS)CommunicationCustomer Relationship Management (CRM)Data StructuresFlaskGen AIGitHub

About

Associate AI Engineer at NPCI, building production-ready Generative AI, NLP, and multi-agent systems that power real UPI workflows at national scale. I specialize in designing LLM-driven automation, graph-based retrieval systems, and secure AI agents that deliver faster response times, higher accuracy, and end-to-end workflow orchestration across NPCI’s payments ecosystem. My work spans multi-agent architectures, Neo4j knowledge graphs, API tool-calling frameworks, fraud detection models, and enterprise AI tools built with strict data-security compliance.

Experience

3 yrs 4 mos

Total Experience

1 yr

Average Tenure

1 yr 8 mos

Current Experience

National payments corporation of india (npci)

2 roles

Associate AI Engineer

Promoted

Oct 2025 – Present · 8 mos · On-site

Engineered a multi-agent LLM system using LangGraph and LangChain for UPI
operations: Q&A assistant for payment queries, mandate tracker for recurring
payments, and grievance platform for transaction disputes, handling 10K+
monthly queries at 92% resolution rate.
Orchestrated hybrid graph retrieval framework with multi-hop Neo4j
traversals, entity boosting, and end-to-end ingestion pipelines, improving
complex UPI query resolution accuracy by 40% with near-zero latency on 50K+
relationship nodes.
Architected hybrid RAG system using Neo4j + Pinecone for UPIHelp Phase-2
that traces payment relationships (user → mandate → merchant →
transaction), delivering 35% more accurate answers to complex multi-entity
queries compared to vector-only search.

Python (Programming Language)Gen AIMCPLangGraphLarge Language Models (LLM)LangChain+2

Graduate Engineer Trainee(GenAI)

Oct 2024 – Oct 2025 · 1 yr · On-site

Designed Pydantic-AI tool-calling infrastructure connecting 15+ UPI APIs for transaction lookup, complaint filing, and mandate status checks, processing 10K+ automated API calls monthly with 45% faster response times.
Created an enterprise-grade ChatGPT for NPCI employees, ensuring 100% compliance with internal data security policies.
Developed Whisper-based AI transcription and summarization system for internal meetings and documentation, accelerating report generation by 70% and reducing manual note-taking effort by 50% across teams.
Built fraud detection ML models for International I-UPI detecting suspicious cross-border payment patterns in real-time, catching 95% of fraudulent transactions while reducing false alarms from 25% to 5% (20% decrease in false positives).
Orchestrated backend services and agent orchestration layer using MCP (Model Context Protocol) to seamlessly integrate mandate management, grievance redressal, and Q&A capabilities into a unified conversational interface.

Python (Programming Language)Machine LearningData AnalysisNatural Language Processing (NLP)Artificial Intelligence (AI)Large Language Models (LLM)+5

Mioneer private limited

Django Intern

Sep 2024 – Oct 2024 · 1 mo · Remote

Created a dynamic garment customization feature using Django, enhancing user experiences and boosting engagement by 30%.
Optimized database management and query handling, resulting in 20% faster response times and improved application performance.
Worked closely with cross-functional teams, contributing to feature enhancements and ensuring seamless integration of new functionalities.

Python (Programming Language)DjangoBitbucketREST APIsGitHubWeb Development

Turing

Microsoft LLM - Data Scientist

Sep 2024 – Oct 2024 · 1 mo · Remote

Python (Programming Language)Large Language Models (LLM)Machine LearningData Science

Futrainc - complete business solutions

AI Developer Intern

Jul 2024 – Aug 2024 · 1 mo · Remote

Developed data preprocessing pipeline applying deduplication, validation, and standardization for LLM training datasets, improving data quality by 25%.
Executed fine-tuning experiments and evaluation runs, optimizing hyperparameters and training data that boosted LLM accuracy by 15% on tasks.
Documenting experiments and collaborating with team members on innovative AI solutions.

Machine LearningLarge Language Models (LLM)Python (Programming Language)FlaskDjango

Celebal technologies

Data Engineer Intern

Jun 2023 – Aug 2023 · 2 mos · Remote

Developed a calculator interface with Python's Tkinter, increasing user engagement by 20% and ensuring accuracy in arithmetic operations.
Engineered a bill management system using JavaScript, boosting user sign-ups by 30% with secure login and account management.
Integrated a MySQL database for secure storage of user credentials, earning Gold Badge recognition for outstanding internship performance.

HadoopProblem SolvingBig DataPython (Programming Language)Apache SparkSQL+1