Skip to content
View ServerCrash358's full-sized avatar
30 followers · 56 following
  • Bengaluru, Karnatake
  • 06:48 (UTC +05:30)

Achievements

Achievement: Pull Sharkx2Achievement: QuickdrawAchievement: Pair ExtraordinaireAchievement: YOLO

Achievements

Achievement: Pull Sharkx2Achievement: QuickdrawAchievement: Pair ExtraordinaireAchievement: YOLO

Highlights

  • Pro

Organizations

@devsper-com

Block or report ServerCrash358

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
ServerCrash358/README.md


About Me

name: Shubhang Srole: Backend & DevOps Engineerfocus:
- Production ML / LLM infrastructure (RAG, multi-agent systems)
- Cloud-native infrastructure, GitOps & DevOps automation
- Distributed systems & backend scalability
- Observability & self-healing infradeveloping:
- advanced ML
- DevOps
- backend
- cloud

I build production-grade ML & backend systems — from RAG and multi-agent LLM infrastructure down to the cloud-native platforms and GitOps pipelines that keep them running reliably.

  • Designing production ML systems — retrieval pipelines, rerankers, and reliability/safety layers for agentic LLM workloads
  • Shipping Kubernetes, GitOps, and IaC workflows with full observability (Prometheus + Grafana)
  • Strong focus on performance, reliability, and self-healing infrastructure
  • Hands-on with scalable backend architectures and automated, verifiable infra pipelines

GitHub Stats


Tech Stack

Languages

Backend & Data


Frontend

DevOps & Cloud



AI / ML


MLOps & Observability


Systems · Hardware · Blockchain


Featured Projects

Mnemosyne - Multi-Agent LLM Safety Layer

A transactional safety layer for multi-agent LLM systems.

  • Deterministic replay & prefix rollback
  • Cryptographic provenance for agent actions
  • Verifier-gated consensus before commit

PythonMulti-AgentLLMReliability

FreshDex - Freshness-SLO-Driven Streaming RAG

A RAG platform built around one invariant: the vector index should never silently drift from the source of truth.

  • Captures document changes via CDC (Debezium reading the Postgres WAL) and streams them through Kafka for near-real-time re-embedding, replacing brittle nightly batch re-indexing
  • Treats index freshness as a first-class SLO — change-to-queryable latency instrumented and burn-rate alerted in Prometheus/Grafana
  • Checksum-gated upserts skip redundant re-embedding; a dependency-aware Redis cache invalidation layer is driven directly by CDC events

RAGDebeziumKafkapgvectorPrometheus

NimbusX - Decentralized Cloud Cost Optimization

Schedules container workloads across compute providers to minimize cost.

  • Blockchain escrow for trustless provider settlement
  • Priority-based workload scheduling
  • Real-time pricing oracles for cost-aware placement

GoBlockchainSchedulingCloud Cost

Lumina-RAG - Production RAG API

Answers questions grounded in your own documents, built to run in prod.

  • Async FastAPI + PostgreSQL/pgvector (HNSW ANN) + Redis cache
  • Two-stage retrieval: vector search then cross-encoder rerank
  • Containerised, deployed to Kubernetes via GitOps, fully observable
  • Automated eval pipeline (MLflow + Prefect) & Terraform on AWS

FastAPIpgvectorRAGKubernetesTerraform

CertOps - Autonomous DevOps Remediation

Simulates, verifies, and certifies infra fixes before they hit prod.

  • Detects CrashLoopBackOff and system anomalies
  • Verifies remediation actions before execution
  • Improves debugging workflows in containerized systems

KubernetesAutomationDevOps

ThreatFind - Autonomous Threat Hunting

Real-time network threat hunting combining vision + retrieval.

  • Vision Transformers (ViT) for anomaly detection
  • RAG pipeline powered by a Llama-based LLM
  • Scalable, security-focused ML inference

PythonViTRAGSecurity


Current Focus

AreaWhat I'm Exploring
Production ML / LLM InfraRAG systems, rerankers, multi-agent LLM reliability & safety
Advanced MLTransformers, ViT, retrieval & evaluation pipelines
DevOps & GitOpsAutonomous remediation, Kubernetes automation at scale
Cloud-Native BackendAsync APIs, distributed systems, AWS + Terraform IaC
ObservabilityMetrics, tracing & self-healing infrastructure

Connect With Me

"Build systems that heal themselves."

Pinned Loading

  1. MnemosyneMnemosynePublic

    A transactional safety layer for multi-agent LLM systems - deterministic replay, cryptographic provenance, verifier - gated consensus, and prefix rollback.

    Python 4

  2. FreshdexFreshdexPublic

    A RAG system built around one invariant: the vector index should never silently drift from the source of truth. Most RAG systems re-index on a batch schedule, so a document can change in Postgres a…

    Python

  3. Lumina-RAGLumina-RAGPublic

    A production RAG API that answers questions grounded in your own documents, built the way it would run in production: async FastAPI, PostgreSQL + pgvector (HNSW ANN), Redis cache, two-stage retriev…

    Python

  4. CertOpsCertOpsPublic

    CertOps is an experimental autonomous DevOps system that simulates, verifies, and certifies infrastructure remediation actions before executing them in production.

    Python 2

  5. NimbusXNimbusXPublic

    A decentralized cloud cost optimization protocol that schedules container workloads across compute providers using blockchain escrow, priority-based scheduling, and real-time pricing oracles.

    Go 1

  6. ThreatFindThreatFindPublic

    ThreatFind is a real-time autonomous network threat hunting system that combines Vision Transformers (ViT) for anomaly detection with a Retrieval-Augmented Generation (RAG) pipeline powered by Llam…

    Python 2