Hire Mounesh H. - RAG Engineer

* I worked as a Python Backend Developer building high-performance microservices, real-time data pipelines, and LLM-powered backend systems for enterprise clients across telecom, banking, and security domains. * Key Contributions * SIEM++ (Vodafone) — Log Analytics & Offense Enrichment Platform * Architected and deployed FastAPI microservices to fetch, parse, and enrich IBM QRadar offenses using LLM-based reasoning. * Designed OpenSearch indexing models enabling sub-second retrieval and analytics. * Performed large-scale syslog EDA, benchmarking Pandas, cuDF, DuckDB, and Polars — delivering up to 3× faster data processing. * Built modular pipelines for offense correlation, rule-based enrichment, and structured insight generation. * Contact Center ASR System (Fino Payments Bank) * Developed a real-time speech-to-text transcription system using FastAPI, open-source ASR models, and scalable chunk-processing logic. * Implemented SIPREC audio ingestion, audio decoding, and streaming architecture. * Delivered live transcription to agent dashboards via WebSocket-based push services. * RAG Backend (State Bank of India – SBI) * Delivered a production-ready Retrieval-Augmented Generation service using document embeddings, vector search, and FastAPI. * Built ingestion pipelines for PDFs, Markdown conversion, chunking, and embedding generation. * Integrated vector DB + OpenSearch to enable intelligent enterprise document search. * Tech stack & Responsibilities * Backend: Python, FastAPI, REST APIs, microservices * Data: Pandas, cuDF, DuckDB, Polars, NumPy * Search/Logs: OpenSearch, QRadar, Syslog processing * AI/LLM: Embeddings, RAG pipelines, ASR, LLM-based enrichment * DevOps: Docker, container-based deployments * Collaboration: Mentored junior developers, conducted code reviews, and managed Git branching & release workflows.

* Built automation systems, data pipelines, and computer vision tooling to eliminate manual workflows and enable scalable data processing across education and AI use cases. * Key Contributions * Student Performance Analytics Engine (Dipstick Project) * Built high-volume Excel data pipelines using Pandas, NumPy, OpenPyXL * Generated multi-level insights (student, section, sub-topic, question-wise) * Automated reporting workflows, reducing manual effort by ~70% * Optimized transformations for large tabular datasets * Computer Vision Dataset Pipeline * Developed video-to-image extraction pipeline for large-scale dataset creation * Built preprocessing workflows for object detection training * Integrated labelImg with auto-labeling using trained models * Reduced manual annotation effort via semi-automated pipelines * Image Processing Toolkit * Built utilities for image compression, bulk preprocessing, and standardization * Improved data quality and reduced storage overhead * Annotation Tool (Tkinter GUI) * Developed GUI-based tool for image labeling with preview and progress tracking * Enabled faster dataset generation for internal teams * Web Data Extraction Pipelines * Built scraping systems using BeautifulSoup and Selenium * Extracted structured data from dynamic sources * Backend: Python * Data: Pandas, NumPy, OpenPyXL, Matplotlib * Automation/UI: Tkinter, Excel automation * Web Scraping: BeautifulSoup, Selenium * Frontend: HTML, CSS, JavaScript * Collaboration: Worked independently on automation tools and data pipelines