24°C 12:06 PM GMT+5:30

Jalaluddin Khan

AI Engineer

Hey, I'm Jalaluddin. I'm building production-ready AI systems with a focus on LLM applications, automation agents, retrieval pipelines, and clean user-first tools for real teams.

I love working at the intersection of intelligence, software, and product, turning messy ideas into scalable assistants, dashboards, workflows, and experiments that feel sharp in the hand.

Brands I have worked with

Vitapure Partner brand logo Partner brand logo Obliq Keyspark

Moments

@AIEngineer @jalaluddinkhan

AI workshop

Research notes

Model desk

Build sprint

Night deploy

Exchange talk

Team day

Work

I started Provizraq Labs Pvt. Ltd. on April 3, 2025, with a simple mission: make education more personalized, accessible, and AI-powered for everyone.

Across this 1+ year journey, I have learned by building across diverse domains: solving legacy network problems, working on full-stack observability platforms, speaking with 50+ stakeholders, and creating systems like realtime course generators, 1-bit language model experiments, source-of-truth platforms, and edge-to-interface digital twins. Along the way, I have also helped early-stage startups move from 0 to 1 with sharper products, stronger systems, and faster execution.

Astrix

Full Stack Observability and Security Platform

Agent-based collector, SSOT, digital twin, NetOps, and automation.

Proviz

AI Powered Personalized Learning

Large multimodal models, symbolic reasoning, and knowledge base systems.

Experiments

Network experimentnetwork
Telecom experimenttelecom
Robot experimentrobot
Education experimenteducation
Data center experimentdata center
Chip experimentchips
Bio experimentbio
Quantum computer experimentquantum

Research & Papers

Blue Model Foundational model research outside conventional autoregressive LLM assumptions; representation learning, compact reasoning, and non-standard training objectives. Quantum Key Distribution Engine Q-to-Q secure transfer model over 200 km fiber with measurable channel loss; key-rate analysis, attenuation, and post-quantum transport design. CVE Remediation Graph RAG-driven vulnerability remediation across heterogeneous device fleets; exploit context, patch mapping, risk ranking, and operator-grade resolution paths. Closed Router Context-switching model router with deterministic policy control; lower-latency routing, capability matching, and private inference orchestration beyond OpenRouter-style brokering. RISC-V Chip Digital Twin Executable digital twin for RISC-V chip behavior; telemetry simulation, edge diagnostics, instruction-path modeling, and hardware-aware observability. Zero-Shot Vision Trainer Alternative visual training regime for generalization without dense labeled datasets; semantic anchors, synthetic supervision, and transfer-time adaptation. Null Agent Swarm Lightweight agentic swarm architecture for constrained environments; stateless coordination, minimal memory pressure, and distributed task decomposition. Bit LM One-bit language model exploration for minimal compute systems; quantized inference, tiny-device deployment, and extreme memory efficiency.

Archive

Portrait of Jalaluddin Khan
Jalaluddin at work
Jalaluddin at work
Jalaluddin at work
Jalaluddin at work
Jalaluddin at work
Jalaluddin at work

Changelog

  1. Launched this interactive AI engineer portfolio with draggable archive pieces and case-study cards.

  2. Built an agentic support system for ticket triage, retrieval, response drafting, and escalation.

  3. Shipped a vision ops monitor for realtime anomaly detection across image and video workflows.

  4. Designed RAG evaluation flows for answer faithfulness, citations, confidence scoring, and model review.

  5. Started documenting AI experiments across networks, telecom, robotics, education, chips, bio, and quantum systems.

Read all work ↗
Email copied