• Senior Forward Deployed Engineer at Searce Inc.
  • Mumbai, India

Hi, I'm Sudhir. I build AI agents and LLM systems on GCP and Kubernetes.

Senior forward deployed engineer across media, BFSI, and telecom. I automate workflows with AI agents, engineer LLM pipelines, RAG, and evaluations, and take them all the way to production.

Email me Download résumé

35%lower MTTR with AI-assisted root-cause analysis

100+microservices migrated from AWS to GCP

1 GB/minmedia streams through multimodal AI pipelines

60%faster developer onboarding on self-service GKE

The stack I work in

  • Google Cloud
  • Kubernetes
  • AWS
  • Terraform
  • Python
  • Go
  • Docker
  • Helm
  • ArgoCD
  • LangGraph
  • FastAPI
  • PostgreSQL
  • Redis
  • GitLab CI
  • OpenTelemetry
  • Grafana
  • Prometheus
  • Kong

Experience

Searce Inc.

Senior Forward Deployed Engineer

Promoted Jul 2026. Previously Cloud & AI Engineer, Apr 2024 to Jun 2026.

Apr 2024 to presentMumbai, India

AI-assisted root-cause analysis

Correlates GCP and AWS telemetry, reducing MTTR by 35%. Services are instrumented with OpenTelemetry collectors feeding Loki and Grafana for metrics, traces, and logs.

  • OpenTelemetry
  • Grafana
  • GCP
  • AWS
  • Loki

AI automation for live news media

A pipeline processing streaming media at 1 GB/min, using multimodal models to cluster and group related content and LLMs to classify and prioritize assets.

  • Multimodal models
  • LLMs
  • Streaming media

BFSI banking-document workflows

Multi-process pipelines per document type: cheques, forms, and policies. Sensitive content is masked with Data Loss Prevention models before processing, then embedded with a multilingual model into a low-latency vector search index.

  • Data Loss Prevention
  • Embeddings
  • Vector search

LLM evaluations as a release gate

Evaluations written across multiple models and modalities, integrated into CI/CD as a pre-release quality gate.

  • LLM evaluations
  • CI/CD

Self-service GKE platform

Production-ready GKE environments on demand, cutting developer onboarding time by 60%. A Go controller for namespaces, quotas, and Helm releases cut manual provisioning by 70%.

  • GKE
  • Kubernetes
  • Go
  • Helm

AWS to GCP migration

Led the migration of 100+ microservices, moving Lambda functions to GKE. Custom Kong Gateway plugins kept downstream services backward compatible throughout the cutover.

  • AWS
  • Lambda
  • GKE
  • Kong

Infrastructure as code and GitOps

Reusable Terraform modules for GKE, networking, and IAM, automated with GitLab CI. Standardized CI/CD across 50+ microservices with ArgoCD ApplicationSets for multi-cluster sync and drift detection.

  • Terraform
  • GitLab CI
  • ArgoCD

Zero-downtime Apigee Edge-to-X cutover

Migrated 120+ API proxies with policies, KVMs, and apps. Built GCP service extensions for auth and rate limiting; authored design docs and delivery templates adopted across teams.

  • GCP
  • Apigee
  • API gateway

Madeline & Co.

Software Engineer Intern

Jan 2024 to Apr 2024Remote

  • Cut initial load times by 30% with Next.js code-splitting. Added Jest tests and lint gates to CI.
  • Built responsive Next.js and React front-end features through code reviews and iterative releases.
  • Next.js
  • React
  • Jest

Projects

Multi-agent RAG platform

KubeDoc AI

A FastAPI gateway with JWT authentication and rate limiting, LangGraph orchestration, pgvector hybrid search, Redis semantic caching, and Celery ingestion. Delivers cited answers over SSE; deployed on Kubernetes with Helm and HPA.

  • FastAPI
  • LangGraph
  • pgvector
  • Redis
  • Celery
  • Kubernetes
  • Helm

Corrective RAG with PII guardrails

CognitoShield

Tokenizes PII at ingestion to keep sensitive values out of the vector index, with bounded agent iterations, pluggable storage backends, and 24 offline tests.

  • PII tokenization
  • Corrective RAG
  • Pluggable storage
  • 24 offline tests

Honourable Mention at the Google Cloud Agent Builders Challenge (APAC 2026)

Technical skills

AI and GenAI

  • LangGraph
  • Vertex AI
  • pgvector
  • LLMs
  • Multimodal models
  • RAG
  • LLM evaluations
  • DeepEval
  • Prompt engineering
  • Google ADK
  • Embeddings
  • Vector search
  • Data Loss Prevention (DLP)
  • PII guardrails

Cloud and platform

  • GCP
  • AWS
  • GKE
  • Kubernetes
  • Docker
  • Helm
  • AWS Lambda
  • Linux
  • Microservices

DevOps and IaC

  • Terraform
  • ArgoCD
  • GitLab CI
  • GitHub Actions
  • Infrastructure as code
  • GitOps
  • CI/CD

Backend and APIs

  • Python
  • Go
  • Bash
  • FastAPI
  • Kong
  • PostgreSQL
  • Redis
  • Celery
  • Next.js
  • React
  • SQL
  • REST APIs
  • API gateway
  • Apigee

Observability

  • OpenTelemetry
  • Prometheus
  • Grafana
  • Loki
  • Monitoring and alerting

Education and certifications

Let's talk.

Building GenAI systems or the platform under them? Email is the fastest way to reach me.

jaiswalsudhir2944@gmail.com
LinkedIn Résumé