Photo of Sandeep

hi sandeep here. đź‘‹

Senior Software Engineer, India 🇮🇳

9+ years shipping search at scale — Solr/Lucene, distributed systems, and ML in production.

🤝 Open to consulting & mentoring

Powered by chai and clean logs. If it vibes, it ships.

LinkedInGitHubEmail
  • Bloomreach

    Bloomreach India

    Senior Software Engineer · Discovery (Search & Recommendations)

    • Architected T5-based ML pipeline (BRIP) processing 10M+ products/day for product categorization.
    • Designed custom Solr ranking with ML optimization, improving relevance by ~30%.
    • Leading semantic search infra design focused on latency and scalability.
    • Collaborating with Staff Engineers on roadmap and ML model integration.
  • Bloomreach

    Bloomreach India

    Software Development Engineer 2 · Discovery

    • Multilingual search using mt-5 models and Solr; extended query relaxation and precision modes.
    • ChatGPT integration with hallucination detection achieving ~95% validation accuracy.
    • Real-time “Recently Viewed” pipeline via Kafka + Spark Streaming (1M+ events/day).
    • Python 2→3 migration across services; infra cost reduced by ~20%.
    • Introduced Istio service mesh for Django microservices; debugging time reduced ~40%.
    • Re-architected Airflow pipelines; migrated to AWS MWAA; SLA improved to ~99.9%.
  • Soroco

    Soroco India

    Senior Software Developer

    • Architected 11 enterprise automation systems for Fortune 500 clients.
    • Led a team of 6 through design reviews and mentoring; team velocity +35%.
    • Designed distributed, event-driven frameworks for high-throughput processing.
  • Soroco

    Soroco India

    Software Developer

    • Built distributed automation systems using microservices and message queues.
    • Championed TDD to ~85% coverage; production incidents down ~60%.
  • MakeMyTrip

    MakeMyTrip

    Android Developer

    • Developed core Android features (filters, sorting, GST) for 10M+ users.
    • Filtering algorithm reduced search time ~50%; MVVM + databinding improved perf ~30%.

featured projects

view more

BRIP ML Pipeline (T5)

Inferred product type extraction at scale (10M+ products/day) to improve product categorization and search relevance.

T5PythonML Pipelines

Multilingual Search (mt-5 + Solr)

Extended advanced search features to non-English languages with mt-5 models and Solr relevance optimizations.

mt-5SolrNLP

Recently Viewed (Kafka + Spark)

Real-time user event ingestion and serving pipeline processing 1M+ events/day with sub-second latency.

KafkaSpark StreamingReal-time

certifications & contributions

  • CodeChef CCDSAP – Advanced DS & Algorithms, Distinction (2017)