s
soban_ahmad_

Soban AI-ML

@soban_ahmad_

AI Automations and ML

Pakistan
English
About me
I run KodoAI, building AI systems. I specialize in identifying manual processes that eat margin and automating them to enhance operational efficiency. My background includes leading engineering teams for SaaS platforms and optimizing high-traffic logistics systems.... Read more

Skills

s
soban_ahmad_
Soban AI-ML
Offline • 
Average response time: 1 hour

See my services

Prospects List Creation
I will build an automated b2b lead gen and outreach system
Full Stack Web Applications
I will develop a secure custom backend, API, and client portal, CRM or erp system

Portfolio

Work experience

AI Systems Engineer

KodoAI • Full-time

Aug 2025 - Present1 yr 1 mo

Lead customer discovery and technical scoping with founders and operations teams; map workflows, define automation boundaries and ship AI systems, API integrations and internal tools from prototype to production. Designed and deployed AR&CO Law Associates' operations platform in eight weeks, replacing a roughly two-hour manual onboarding process with an under-three-minute client flow used by 10-15 staff. Built 12+ modules across 20+ PostgreSQL tables for intake, case numbering, booking, secure document exchange, WhatsApp communication, invoicing and live reporting using Next.js, NestJS, Supabase/PostgreSQL and AWS. Implemented exact CNIC pre-screening, validation gates and service-based routing while retaining lawyer approval or rejection for judgment-sensitive decisions.

Systems_Limited

PK Delivery Management

Systems Limited • Full-time

May 2025 - Aug 20253 mos

Built NestJS microservices and PostgreSQL APIs, cutting latency by 35% for a live delivery-management system; developed tracking for 150+ deliveries per day and dashboards that improved internal workflow efficiency by 20%.

Axiss

Full-Stack & QA

Axiss • Full-time

Dec 2024 - May 20255 mos

Built a generative-AI chatbot over REST APIs that increased user engagement by 25%; tightened QA and debugging workflows, reducing delivery time by 15% and debugging time by 20%