3 viewsjobseeker
Susanth N. — Mid-Level AI Systems Engineer from India

Susanth N.

Mid-Level AI Systems Engineer

India 3-6 years
Open to offersNew to Platform
Languages
EnglishHindi
Video Introduction
No video introduction yet
The candidate has not added a video.
Contact information and social networks are private. Connect to unlock.
Hidden

About

Susanth N. is an accomplished AI Engineer specializing in the architecture and deployment of production-grade AI systems. With over three years of experience in Agentic AI Systems and Platform Engineering, Susanth is committed to building scalable and reliable LLM-powered applications. At DANZ in Hyderabad, he has developed a sophisticated RAG-based developer assistant, serving over 800 engineers, which indexes extensive internal documentation to streamline developer workflows. His work includes the integration of hybrid retrieval systems, incorporating BM25 and dense vector search to enhance query precision. Susanth has also excelled in implementing secure AI functionalities such as RBAC access control and structured output validation. His deployment strategies leverage cloud-native infrastructure technologies like Kubernetes, ensuring efficient lifecycle management of AI workloads. His efforts have significantly reduced information discovery times and improved both cost efficiency and response accuracy. Susanth is adept in handling DevOps tasks, further improving deployment cycles and infrastructure management in cloud environments.

Experience

  • AI Engineer – Agentic Systems (Platform Engineering)

    DANZ · 2023 — Present
    Developing a developer assistant utilizing a RAG-based architecture to support approximately 800 engineers by indexing over 25,000 API documents and changelogs. Creating a multi-step agentic workflow that includes query classification, dynamic retrieval strategy selection, and response validation. Implementing a hybrid retrieval system that combines BM25 (Elasticsearch) and dense vector search (Qdrant/pgvector) with RRF fusion, alongside training a query classification model based on more than 500 internal inquiries. Designing mechanisms for format-aware chunking that preserve code block integrity and handle version-aware changelogs. Developing a LangGraph-based agent to automate incident triage, incorporating structured debugging steps and human-in-the-loop validation. Establishing security measures for prompt management and structured output validation, alongside integrating RBAC access controls and evaluation pipelines.
  • Reliability, Performance & Observability Engineering

    DANZ · 2023 — Present
    Managing error handling mechanisms for graceful system degradation via alternate retrieval methods. Identifying and addressing failure modes like stale retrieval and latency issues in vector databases. Deploying circuit breaker patterns for LLM APIs to mitigate cascading failures. Instrumenting observability throughout pipeline stages and setting up real-time monitoring dashboards to analyze retrieval quality and failure metrics. Optimizing latency through caching strategies across query and response layers using Redis.
  • Cloud Infrastructure, DevOps & Platform Engineering

    DANZ · 2023 — Present
    Overseeing a 20-node EKS cluster supporting over 60 microservices while maintaining on-call ownership. Enhancing CI/CD pipelines through the addition of automated testing and security checks. Developing high-throughput asynchronous processing with SQS and Celery, ensuring idempotency for reliable operations. Automating infrastructure provisioning with Terraform across multiple AWS accounts and implementing cross-region strategies for improved service availability. Standardizing access control in Kubernetes and advancing observability with tools like Prometheus and Grafana.

Skills & Expertise

Education

  • B.A.
    Vinayak Mission Sikkim University · 2016 — 2019