Shubham Pandey
Software Engineer exploring AI frontiers | New to tech blogging but eager to share insights | Always learning, always connecting
shubham.pandey.workconnect@gmail.comBut, who am I really? ->
Ciena
Software Engineer 1B - RA Team
- •Currently building Resource Adapters that enable major networking companies to efficiently manage their OSS (Operations Support Systems) suites.
- •Architected and implemented an end-to-end Resource Adapter that interfaces with customer devices via CLI, transforming raw responses into structured resource models aligned with Ciena BluePlanet Orchestrate specifications. Streamlined resource provisioning workflows through robust object-oriented design patterns, ensuring scalability and maintainability. Delivered high code quality through comprehensive unit test coverage and leveraged AI-assisted development tools to accelerate delivery timelines by 30%.
- •Designed and implemented a logical inventory system with Forwarding Relationship Encapsulation (FRE) to accurately map physical and logical Terminating Point Encapsulations (TPEs). Extended orchestration capabilities through precise relationship modeling, enabling streamlined management of complex network topologies using object-oriented design principles.
Mercor
Software Engineer (Contractor)
- •Standardized Infrastructure: Automated the containerization and environment parity of 50+ high-traffic Python repositories using Docker, facilitating high-fidelity data acquisition and reproducibility for a leading AI research lab.
- •Evaluation Engineering: Architected atomic rubrics for complex System Design problems, establishing a rigorous benchmarking framework used to evaluate the architectural reasoning of AI models.
- •Technical Leadership: Directed rigorous peer code reviews focused on algorithmic efficiency and edge-case detection, maintaining production-grade code quality across distributed codebases.
Hughes Systique
Associate Software Engineer - AI Team
- •ML Document Pipeline: Engineered automated processing system using Power Automate and Python, extracting 10,000+ monthly vendor documents with 95% accuracy, reducing processing time from 4 hours to 30 minutes
- •Data Architecture Design: Architected serverless pipeline using GCP Cloud Run and BigQuery, processing 100+ vendors’ records and fault data daily, enabling real-time analytics
- •SSO Access Control: Implemented OKTA SSO authentication using OAuth 2.0 for DMS, securing vendor document access and processing for external partners with zero security incidents
- •AI Assistant: Developed full-stack AI assistant using NextJS and FastAPI, driving 50%+ adoption across 1000+ employees with a data-driven dashboard for usage analytics and continuous feature deployment
- •Backend Infrastructure: Architected high-performance FastAPI backend with robust exception handling, reducing response time by 50%, utilizing MySQL and Weaviate, and containerizing with Docker for scalable deployment.
- •Langchain Framework: Engineered custom functionality by overriding framework entity classes, implementing an advanced Reranker Model for document embeddings that significantly enhanced search accuracy by 80%.
- •Testing & Deployment: Prepared comprehensive test strategy for LLM-powered chatbot, leveraging GPT-4 to automate 80% of manual testing, and implemented CI/CD workflows using JIRA to enhance quality assurance and development efficiency. Used Docker for Continuous deployment to production server
Hughes Systique
Software Engineer Intern - AI Team
- •Developed a Customer Support Assistant prototype for a leading cruise line company. The solution involved web scraping using Selenium and Beautiful Soup to gather data and documents, which powered a RAG-based AI assistant. Enhanced the system with guardrails and semantic analysis to ensure reliable, contextually accurate responses.
Education
MCA — Master of Computer Applications
Jawaharlal Nehru University (JNU)
- •CGPA: 8.47/10
Bachelor's — Computer Science
University of Delhi
- •CGPA: 8.4/10