NameAneesh Aparajit

Experience

Where I have been and what I built there. Mostly backend, sometimes a search index, occasionally a model, always a retry policy.

Roles

  1. Zeta

    Jul 2026 – Present

    Software Engineer IIBengaluru

    Co-designed the authorization gateway for savings-account bill payments, bridging our ledger and a partner bank's, and built idempotent reconciliation and regulatory reporting pipelines on Apache Flink with Camunda workflows.

  2. Zeta

    Nov 2024 – Jul 2026

    Software Engineer IBengaluru

    Owned the KYC and merchant modules, redesigning KYC around PAN, bank-account and video verification, and shipped fraud controls such as inbound OTP checks and card cool-off limits.

  3. Trademarkia

    Jan 2024 – Nov 2024

    Software EngineerChennai

    Built a document-signing platform in Go and Java with role-based access, and logo and court-case search on Elasticsearch and Qdrant fed by automated ingestion jobs.

  4. Samsung Research Institute Bangalore

    Jan 2024 – Nov 2024

    Software Engineer, PRISM (part-time)Remote

    Sped up model training with LoRA fine-tuning to make video processing practical, and built internal web and desktop tools for the research team.

Skills

Languages
Go, Java, Python, C++, SQL
Backend
Spring Boot, Fiber, FastAPI, gRPC, REST APIs, Apache Kafka, Apache Flink, Camunda
Databases
PostgreSQL, Redis, MongoDB, Elasticsearch, Qdrant, Pinecone
Cloud and DevOps
AWS (S3, Lambda), Docker, Jenkins, Prometheus, Grafana
Machine learning
PyTorch, Transformers, LoRA/PEFT fine-tuning

Education

  1. Vellore Institute of Technology

    Sep 2020 – Jun 2024

    B.Tech in Computer ScienceChennai

Achievements

  • Zeta Stellar Award for 2026, along with 4 Spot Awards and 5 Ultimate Team Awards.
  • First Runner-Up, NASSCOM Telangana AI Mission (TAIM) 2023, out of 7,500+ teams; Finalist, Bajaj HackRx 4.0 (top 36).
  • Excellence Award for dichotomous image segmentation, Samsung PRISM.
  • Open-source Python packages reddit-multimodal-crawler (9,100+ downloads) and torchtemplates (3,700+ downloads).
  • 2x Kaggle Expert: top 0.6% in Datasets and top 0.7% in Notebooks.