Hello, I'm
Zishan Khan
Backend and AI Systems Engineer architecting production multi-agent runtimes, real-time Voice AI pipelines, and distributed backends with verified sub-150ms inference latency.
Zishan Khan
AI Systems & Backend
Architecture & Engineering
Production Backends · Multi-Agent AI · Real-Time Systems
I am a Backend and AI Systems Engineer focused on building production multi-agent architectures, real-time Voice AI runtimes, and high-throughput vector search pipelines.
From deploying mission-critical government infrastructure at THDC India across 9 modules and 91 source files to engineering autonomous Voice AI telephony runtimes with sub-second interruption handling, I design software that is resilient, audited, and optimized for real-world production workloads.
Multi-Agent Systems
LangGraph orchestration, BullMQ queues, and autonomous supervisor-critic loops.
Voice AI & Telephony
Twilio streaming ASR, ElevenLabs neural TTS, and sub-second interruption handling.
Vector & RAG Memory
PostgreSQL + pgvector, Redis intent caching, and verified sub-150ms p95 latency.
Production Backends
Node.js, TypeScript, Express, Fastify, multi-tier RBAC, and Docker deployments.
Agentic & Telephony Ecosystem
Production Autonomous Runtimes
Production multi-agent architectures, real-time Voice AI runtimes, and high-throughput vector retrieval pipelines engineered for autonomous decision-making and low-latency execution.
Autonomous Voice & Multi-Agent Workflows
Building production telephony and agentic pipelines that integrate Twilio streaming ASR, LangGraph state machines, and ElevenLabs neural TTS. Backed by Redis/BullMQ task queues and pgvector memory, these systems handle live calls, interruption events, and structured tool execution.
Vector Search
pgvector + Redis
Task Queues
Redis + BullMQ
Voice Telephony
Twilio + ElevenLabs
Orchestration
LangGraph State
AI Appointment Assistant
Voice-powered · Appointment AI · Male voice
Full AI Stack
Skills & Frameworks
Agentic Workflows
LangGraph · LangChain
Advanced RAG
pgvector · Hybrid Retrieval
Context Engineering
Stateful Dynamic Assembly
LLM Evals
Benchmarking · Metric Scores
AI Observability
Prometheus · MLflow · Grafana
Guardrails & Safety
Schema Validation · NeMo
Voice AI & Telephony
Twilio · ElevenLabs · Audio ASR
Semantic Search
Embeddings · Cross-Encoder
AI Task Queues
Redis · BullMQ Workers
Multi-Agent Systems
Supervisor · Worker DAGs
Model Fine-tuning
LoRA · PyTorch Pipelines
Tool-Use & MCP
Model Context Protocol · APIs
End-to-End Automation
Designing pipelines from data ingestion → reasoning → action → feedback loops
Featured Systems & Engineering
Production-grade multi-agent architectures, real-time Voice AI runtimes, and distributed diagnostic pipelines engineered for high throughput and verified reliability.
Interactive AI / ML Workspace & Neural Lab
Experience client-side deep learning in action: Run real-time 1D CNN telemetry classification, explore Apriori association graphs, and simulate epoch training directly in your browser.
Haanna – Production RAG Platform
Production RAG companion platform serving 200+ concurrent users with pgvector persistent semantic memory and sub-150ms p95 latency.
Platform.voice – Autonomous Voice AI & Business Agents
Real-time autonomous Voice AI telephony platform integrating Twilio, streaming ASR, LangGraph multi-agent orchestration, and ElevenLabs neural TTS.
AI Learning Path Generator – Agentic System
Production agentic learning platform orchestrating Composio OAuth tools to auto-generate day-wise curriculums and write to external productivity suites.
Intelligent DCRM Diagnostics & Telemetry Pipeline
Real-time analytics engine processing telemetry streams for automated fault detection and diagnostic reporting.
adClickMagnet
Interactive marketing and web engineering platform featuring 3D Spline experiences and high-performance GSAP animations.
Experience & Education
My professional journey so far.
Backend Developer
THDC India Limited
Architected and deployed a production REST API backend in Node.js/TypeScript on government infrastructure across 9 modules and 91 source files. Designed multi-tier RBAC with audit trails (>80% Jest coverage via GitHub Actions CI) and Cloudflare R2 document pipelines.
Freelance Software Engineer
Self-Employed (Remote)
Engineered Haanna, a production RAG platform serving 200+ concurrent users with pgvector persistent memory (<150ms p95 latency), Socket.io real-time streaming, Redis-backed BullMQ queues, and Prometheus/Grafana monitoring.
Full-Stack Developer
KKRF Group
Built Next.js web applications with Three.js 3D components, GSAP animations, and Spline integrations. Owned full project lifecycle across three client engagements with Lighthouse-optimized performance and SEO.
Full-Stack Developer
Bigbooster Technology
Shipped full-stack SaaS features using Next.js, TypeScript, and AWS S3. Built a reusable Tailwind CSS component library that reduced UI development time across multiple production screens.
Technical Trainer
Archway
Delivered backend and web development workshops; mentored aspiring engineers through hands-on project-based learning.
B.Tech in Computer Science & Engineering
IPS Academy (IES), Indore
Focusing on distributed systems, machine learning engineering, and scalable backend architecture.
Awards & Recognition
Honored to have my work and dedication recognized on national platforms.