H
i
,
I
'
m
D
a
w
o
o
d
.
I'm an AI Engineer focused on LLM systems, voice agents, RAG pipelines, and production Full Stack Systems at Kodensoft
A
b
o
u
t
M
e
Hi! I am Dawood Shahzad, an AI and backend engineer based in Karachi, Pakistan. My strongest work sits where applied AI meets reliable backend systems: LLM apps, RAG assistants, voice agents, automation workflows, and real-time APIs.
I have a Bachelor's degree in Artificial Intelligence and 2+ years of hands-on experience building startup products with Python, FastAPI, OpenAI, LangChain, RAG, PostgreSQL, Redis, Docker, and cloud deployment workflows. I care about systems that are fast, accurate, maintainable, and practical in production.
Recently I have worked across AI-enabled web, mobile, backend, and automation products at Kodensoft, Nosh Solutions, RealBotAI, and Authentik. I also build frontend and mobile experiments when the product needs a complete interface, but backend and AI/ML are my center of gravity.
When I am not working, I like to spend time with anime, AI novels, and scientific documentaries.
E
x
p
e
r
i
e
n
c
e
Full Stack Developer
>Full Stack Developer @ Kodensoft
February 2026 - Present
- Built AI-powered web, mobile, and backend features across Kallin.ai, OhCrema, and AmeriDistributors.
- Shipped FastAPI services, database-backed APIs, integrations, and deployment support for startup product workflows.
- Led work on a Flutter e-commerce application and helped integrate an AI agent for automated food ordering.
F
e
a
t
u
r
e
d
P
r
o
j
e
c
t
s
Voice AI Backend
Custom Voice Agent and Backend
Built a Pipecat-based phone-ordering agent with FastAPI, Twilio, Redis, and OpenAI. The system reduced multi-query delay from 5-10 seconds to about 300 ms using in-memory hash tables, Redis-backed cart queries, fuzzy search, phonetic matching, Levenshtein distance, and transcription-aware auto-correction while cutting projected platform cost by about 80%.
FastAPI
OpenAI
Twilio
Redis
Voice AI
Backend
AI Computer Vision Product
AuraTry - AI Virtual Try-On
Fine-tuned SDXL with LoRA for an AI virtual try-on workflow and built the supporting FastAPI backend and mobile experience. The project combines model customization, image generation pipelines, API design, and mobile product integration for a practical fashion-tech use case.
SDXL
LoRA
PyTorch
FastAPI
React Native
AWS
AI Automation System
Voice Agent Lead Outreach
Built a voice follow-up system using VAPI, n8n, GoHighLevel, and webhook-driven CRM automation. The workflow increased lead generation by 40% and automated call logging, status updates, follow-up triggers, and lead qualification handoffs.
VAPI
n8n
GoHighLevel
Twilio
Webhooks
CRM Automation
RAG Chatbot Platform
LuxeBot AI
Built an AI chatbot platform for automated lead qualification, appointment scheduling, and 24/7 client support. The system used LangChain and RAG patterns to reduce response time by 75% while integrating with CRM workflows for agency operations.
Clinical RAG Assistant
Novik - AI Dental Assistant
Created a GPT-4 and FAISS based assistant for dental professionals, handling more than 500 queries per day with 95% accuracy. Novik supports evidence-aware recommendations across preoperative precautions, anesthetics, antibiotics, analgesics, postoperative care, and patient-specific context.