Skip to main content
Sonu Yadav
  • Home
  • About
  • Work
  • Projects
  • Blog

© 2026 Sonu Yadav · Labs of Universe. All rights reserved.

  • LinkedIn
  • GitHub
  • Email
  • RSS

Projects

Selected projects where I led or heavily contributed to system design and implementation—architecture, stack, and outcomes.

  • Traffic Analytics (Trafficure)

    2024 – Present

    Client: Cities / Municipal

    End-to-end architecture for city-level traffic analytics: high-throughput ingestion, geospatial processing, and dashboards that help reduce congestion and logistics cost.

    PythonPostgreSQLPostGISDockerAWSETL
  • Address Matching & Validation

    2024 – Present

    Client: Multiple enterprise clients

    Architected a configurable address-matching pipeline for Indian addresses at scale—normalize, match, score—reducing manual verification and feasibility checks for enterprise clients.

    PythonPostgreSQLFastAPIDockerLinux
  • Location Intelligence for Fuel Stations

    2024 – Present

    Client: IOCL

    System design and implementation for IOCL’s location intelligence: geospatial analytics, multi-layer data integration, and decision support for siting new fuel stations across India.

    PostgreSQLPostGISPythonGISDocker
  • 360° Video Processing Pipeline

    2024 – 2025

    Designed and led a staged, cost-efficient microservices pipeline for 360° video; reduced processing cost by ~30% through architecture and resource optimization.

    PythonDockerAWSMicroservices
  • Telecom ETL & KPI Calculation

    2024 – 2025

    Client: Tata Communications, Vodafone Idea

    Architected ETL and KPI pipelines for large-scale telecom data; owned schema design, job orchestration, and SLA for client-facing analytics (Tata Communications, Vodafone Idea).

    PythonPostgreSQLETLDockerLinux
  • AI Customer Support Assistant

    Jul – Sep 2023

    Client: GajiGesa

    Designed and built a chat-based AI assistant (Django, OpenAI, Langchain) with clear integration points and failure handling; reduced ops workload by ~3 hours per person daily.

    PythonDjangoOpenAILangchain