Profile

AI Engineer specializing in production-grade LLM pipelines, multi-agent workflows, and advanced RAG architectures. Expertise in combining vector databases with ontology-driven knowledge graphs to build highly accurate, context-aware AI systems across complex enterprise data.

Professional Experience

Full Stack LLM Development Associate

Accenture
09/2024 – Present | Gurugram, India
  • •Engineered a central LangGraph orchestration layer in Python, integrating Pub/Sub for event-driven system triggers and job dependency tracking to automate the execution of structured and unstructured data pipelines
  • •Implemented pipeline observability using Langfuse for distributed tracing, token tracking, and real-time alerting to enable automated failure recovery and seamless debugging.
  • •Containerized and deployed pipeline components as Google Cloud Run jobs, enabling on-demand manual execution and isolated testing of individual pipeline stages outside scheduled LangGraph runs.
  • •Built a Classification Agent to categorize BigQuery table columns by sensitivity, data category, regulatory tags with 95%+ classification accuracy.
  • •Developed AI agents to automate data product generation, featuring a Source Identification Agent that analyzes source data to identify logical data products and their underlying tables, and a Data Mapping Agent to define schema transformation rules.
  • •Engineered end-to-end downstream execution by deploying a Data Product Creation Agent to dynamically generate SQL queries for table creation, and a Hydrate Agent to execute data population.
  • •Created a Metadata Agent to automatically generate comprehensive business and technical metadata for the new data products, including detailed data lineage and descriptions.
  • •Programmed a deterministic Data Quality Agent to strictly enforce BigQuery schema conformance, domain validity, and primary key integrity via automated unit tests.
  • •Engineered and optimized a high-throughput GCP pipeline to stream and process 7,000+ enterprise documents from GCS, applying semantic chunking and embedding models to index vector representations into Elasticsearch maintaining 100% data integrity via BigQuery checkpointing for fault-tolerant execution. 
  • •Engineered a PPT and PDF Extraction Agent leveraging vision-based LLMs to programmatically extract structured text and complex tables from presentation documents stored in GCS, converting raw visual layouts into standardized formats for downstream vectorization.
  • •Created an Entity Extraction Agent to map retrieved vector context to RDF-compliant ontologies, automatically populating an enterprise knowledge graph in Ontotext GraphDB.
    • Skills
    Programming Languages: C , C++ , Python , JavaScript , HTML5 , CSS
    AI, LLMOps & Vector Search: LangGraph, Langchain, AutoGen, Langfuse, Elasticsearch (Vector DB)
    Libraries & Tools: FastAPI, Scikit-learn, TensorFlow, Numpy, Pandas, NLP, OpenCV, Git, GitHub, VS Code, Postman, Tableau
    Data & Knowledge Graphs: SQL, MongoDB, Ontotext GraphDB
    Cloud & Infrastructure: Google Cloud Platform (GCP), BigQuery, Cloud Storage (GCS), Cloud Run, Pub/Sub
    Personal Projects

    JobReady – AI-Powered Job & Resume Analyzer⁠

    Streamlit - Python - OpenAI GPT-4
  • •Developed a Streamlit-based application using GPT-4 to analyze job descriptions and resumes for ATS optimization.
  • •Engineered a core matching engine that performs automated skill gap analysis, keyword extraction, and experience alignment.
  • •Delivers granular match reports, quantifying candidate fit and providing actionable feedback to improve application success.

    • PostCraft AI – AI-Powered LinkedIn Post Generator⁠

    LangChain - Python - Streamlit - Groq API
  • •Built an AI-powered content generator using LangChain , leveraging few-shot learning to dynamically match user tone, featuring adjustable parameters for topic, language (English/Hinglish), and length constraints.
  • •Engineered a robust ingestion pipeline to process raw social media data, automating the extraction and structuring of metadata to enable high-speed, quality-assured data retrieval.

    • Education

    Bachelor In Technology - CSE

    Jaypee Institute Of Information Technology
    2020 – 2024 | Noida, India