Aditya
Available for opportunities

Hi, I'm Aditya Generative AI / Agentic AI Engineer|

3+ years at Accenture delivering production AI systems for Fortune 10/100 clients across healthcare, aviation, and entertainment. Designed a multi-agent failure triage system cutting resolution from 1-2 days to under 10 minutes; architected a real-time healthcare call assistant processing 5,000+ live utterances daily with sub-second latency.

0+ Years Experience
0 GenAI Projects
0 Certifications
0 Fortune Clients
View Projects Get in Touch Resume GitHub
Scroll to explore
// about me

Building Intelligent Systems
That Actually Ship

I'm a Generative AI / Agentic AI Engineer who believes in building AI systems that go beyond demos. With 3+ years at Accenture, I've shipped production AI for some of the world's largest companies.

My work spans the full GenAI lifecycle: from designing RAG pipelines with hybrid search and reranking, to orchestrating multi-agent systems with Semantic Kernel and LangGraph, to fine-tuning LLMs with QLoRA, and building safety guardrails with drift detection and A/B testing.

Beyond my enterprise work, I've built 8 open-source GenAI projects covering every major area of the GenAI stack — RAG, multi-agent systems, fine-tuning, multimodal, prompt engineering, safety/observability, and MLOps — each with comprehensive testing and production-ready architecture.

Enterprise Scale Fortune 10/100 clients
Full Lifecycle Data prep to monitoring
Production Focus Safety, testing, CI/CD
aditya@genai ~
$ cat profile.json 
{
  "name": "Aditya",
  "role": "Generative AI / Agentic AI Engineer",
  "company": "Accenture",
  "experience": "3+ years",
  "location": "Noida, India",
  "domains": [
    "Healthcare",
    "Aviation",
    "Entertainment"
  ],
  "focus": [
    "RAG Pipelines",
    "Multi-Agent Orchestration",
    "LLM Fine-Tuning",
    "AI Safety & Guardrails",
    "Production MLOps"
  ],
  "open_to": "GenAI opportunities"
}
// experience

Professional Journey

Software Engineer - Generative AI

Accenture Solutions Private Limited
Jul 2024 - Present
  • Designed and deployed a multi-agent AI system for a global airline technology provider using Azure AI Foundry and Semantic Kernel — automating failure triage, test generation, and regression scoping, reducing resolution time from 1-2 days to under 10 minutes
  • Built Hybrid RAG pipelines with deduplication and human-in-the-loop workflows, serving 50+ internal users with 90%+ accuracy and sub-60s response, reducing L2 escalations by ~35%
  • Engineered agentic test case generation with regression impact prediction and risk scoring (1-5), processing 500+ daily test scenarios via automated data ingestion into Cosmos DB
  • Collaborated with cross-functional teams across engineering, QA, and product to align GenAI solutions with enterprise workflows and delivery milestones
Azure AI Foundry Semantic Kernel Cosmos DB RAG Multi-Agent

Associate Software Engineer

Accenture Solutions Private Limited
Jan 2023 - Jun 2024
  • Architected a real-time AI assistant for a Fortune 10 healthcare company processing 5,000+ live call utterances daily using Flask, Azure OpenAI, and Cognitive Search with sub-second latency, supporting 200+ agents
  • Built production APIs for NLP identity verification, real-time intent detection (CLU + GPT fallback), confidence-scored auto-suggestions, and post-call summarization — handling 10K+ API calls/day with 99.5% uptime
  • Designed event-driven pipeline (Event Hub → Functions → Cosmos DB → Flask → React) with multi-threaded execution and feedback loops, reducing end-to-end processing latency by 40%
  • Delivered a large-scale GenAI application for a Fortune 100 entertainment company using GPT-4o — automating document generation, competency mapping, and interview guides, reducing manual effort by ~70-80% across 15+ departments
  • Engineered specialized prompt workflows with context injection and inclusive language detection across 8 bias categories, achieving 95%+ compliance on document validation
Azure OpenAI Flask Event Hub Cognitive Search GPT-4o
// projects

Open Source GenAI Projects

8 production-ready projects covering the complete Generative AI stack

DocuMind

Production RAG Pipeline

Production RAG pipeline with 3 chunking strategies, hybrid search (dense + BM25), Cohere reranking, and multi-LLM cited answer generation. RAGAS-style evaluation with 6 preset experiment configurations and interactive results dashboard.

Python OpenAI ChromaDB Pinecone FastAPI Docker
Hybrid Search + Reranking
RAGAS Evaluation

AgentForge

Multi-Agent Research System

5-agent research pipeline with reflection loops in both LangGraph and CrewAI to benchmark framework trade-offs. Integrated guardrails for prompt injection, PII masking, hallucination scoring, and loop prevention.

Python LangGraph CrewAI Tavily ChromaDB Streamlit
LangGraph + CrewAI
Safety Guardrails

ModelForge

LLM Fine-Tuning Pipeline

Fine-tuned Mistral 7B using QLoRA on 500+ code review examples with SHA256 and MinHash deduplication. Evaluation dashboard comparing 4 models across BLEU, ROUGE, BERTScore, and LLM-as-Judge with cost breakeven analysis.

Python HuggingFace QLoRA PEFT BERTScore Streamlit
QLoRA Fine-Tuning
4-Model Comparison

GuardianAI

LLM Safety & Observability Platform

Decorator-based LLM tracing capturing latency, tokens, cost, and 4 quality metrics across 9 model pricing tiers. Implemented drift detection, configurable alert rules, LRU cache, and A/B testing framework.

Python SQLite Streamlit Plotly
LLM Observability
A/B Testing

ShipAI

ML Deployment & DevOps Pipeline

CI/CD pipeline with GitHub Actions, multi-stage Docker builds, and security scanning for dependencies and Dockerfiles. 3-tier health checks and deployment runbooks with step-by-step rollback procedures.

Python Docker GitHub Actions FastAPI CI/CD
Security Scanning
CI/CD Pipeline

VisionRAG

Multimodal RAG Pipeline

Multimodal RAG handling PDF + image ingestion with OCR extraction, vision embeddings, hybrid retrieval, and multimodal answer generation. Processes complex documents beyond plain text.

Python OCR Vision Embeddings Hybrid Retrieval

InterviewGPT

Adaptive Mock Interview System

AI mock interview system with adaptive difficulty, 14 heuristic scoring functions, session management, and progress analytics for personalized improvement.

Python OpenAI Streamlit Analytics

ToolBridge

Prompt Engineering Toolkit

Prompt engineering toolkit with template management, chain-of-thought patterns, few-shot optimization, A/B testing, and version control for prompt iterations.

Python Prompt Engineering A/B Testing Versioning
// skills

Technical Arsenal

GenAI & LLMs

OpenAI GPT-4o Azure OpenAI Service Anthropic Claude Mistral 7B HuggingFace Prompt Engineering CoT / Few-Shot / ReAct RAG Multi-Agent Systems Agentic AI Fine-Tuning (QLoRA, LoRA, PEFT) RLHF SFTTrainer Multimodal RAG Embeddings

Agentic AI & Orchestration

LangChain LangGraph CrewAI Semantic Kernel Azure AI Foundry Multi-Agent Orchestration Tool Calling Function Calling MCP Servers ReAct Pattern Human-in-the-Loop

RAG & Vector Search

Azure AI Search Azure Cognitive Search Pinecone ChromaDB BM25 Hybrid Search Cohere Reranking Vector Embeddings Semantic Search RRF

AI Safety, Evaluation & MLOps

RAGAS BERTScore BLEU / ROUGE LLM-as-Judge Guardrails Content Filtering PII Detection Prompt Injection Prevention Model Monitoring Drift Detection A/B Testing LLMOps

Cloud & Infrastructure

Microsoft Azure Azure Functions App Service Event Hub Logic Apps Blob Storage Key Vault Docker Docker Compose FastAPI Health Checks Security Scanning GitHub Actions CI/CD Azure DevOps

Backend & Data

Python SQL Flask FastAPI Cosmos DB Azure SQL SQLite Streamlit Gradio Pydantic pytest Git REST APIs Microservices Event-Driven Architecture
// certifications

Certified Expertise

Azure AI Engineer

AI-102

Azure Developer

AZ-204

Azure AI Fundamentals

AI-900

AWS Cloud Practitioner

CLF-C02

Azure Fundamentals

AZ-900

MongoDB Certified

Developer
// achievements

Recognition & Awards

2024

Sparkling Star Award

Recognized for building multi-agent GenAI system adopted across enterprise workflows — Accenture & Avanade

2023

ACE Award (Team)

Awarded for building and deploying a real-time healthcare AI assistant handling 5,000+ daily utterances with sub-second latency and 99.5% uptime

// contact

Let's Build Together

Open to GenAI opportunities. Let's connect and discuss how I can contribute to your team.

Email

aditya78332412@gmail.com

LinkedIn

adityasanjay1999

GitHub

aditya2425

Phone

+91-7833934588