Home Original page

archit15singh - Overview

Archit Singh

Senior Backend Engineer | Infra & System Design @ Scale

Kafka · Redis · Python · Postgres · GenAI · Distributed Systems · Observability

whoami

I build backend systems that remain reliable at scale, are observable under failure, and are optimized for real-time detection, memory, and search. Over 8 years across infra-heavy teams, I’ve shipped telemetry pipelines, distributed orchestrators, and GenAI-backed systems under concurrency, latency, and audit constraints.

🔩 What I Build

  • Distributed Cloud Applications → predictable scale, safe concurrency, recoverable failure modes
  • Stream Processing Pipelines → Kafka + Redis + Postgres under 10M+ monthly events
  • Telemetry & Observability Systems → OpenTelemetry, Prometheus, SLA dashboards
  • LLM Agent Infrastructure → memory-augmented, tool-using multi-agent systems
  • Control Planes & Coordination → consensus, retries, failover, eventual consistency

🧠 Core Expertise

  • Distributed Systems: queues, consensus, state machines, async orchestration
  • Infra Design: ingestion, feature stores, detection pipelines, error budgets
  • Memory & Search: Redis state machines, Postgres/pgvector, vector search, explainability
  • Observability: tracing, metrics, SLA diagnostics, chaos/fault injection
  • GenAI Integration: structured planning, retrieval, evaluation pipelines
  • Cloud & Ops: Docker, AWS (ECS, CloudWatch, S3), Terraform, Kubernetes

🚀 Key Outcomes

  • Designed streaming pipelines handling 10M+ events/month
  • Reduced cross-region failures by 35% through retry-safe orchestration
  • Cut ETL latency by 30% in telemetry-heavy clinical systems
  • Built entity-scoring detection pipelines surfacing explainable signals for ML models
  • Logged full agent memory + tool usage telemetry for enterprise GenAI workflows
  • Early Redis-based observability platform → acquired by Redis Inc (folded into RedisInsight)

🛠️ Featured Projects

🧠 memori Open-source, production-grade memory substrate for AI systems. Append-only log in Postgres, fast semantic recall with pgvector, short-term caching in Redis, and lightweight knowledge graph for provenance. Includes reflection jobs, PII redaction, and right-to-delete policies, because memory should be infrastructure.

🌍 Connect with Me

Currently exploring Senior/Staff roles in AI Infrastructure, Detection & Entity Scoring, Memory/Search Systems, or Distributed Systems/Observability. Let’s build resilient, explainable, and future-proof infra.