Let's Talk Resume
Hi, my name is

SYED ASJAD ABBAS

Organization
AI/ML Engineer @ SVYPE · Founder @ Zeploy Tech
Syed Asjad Abbas, AI/ML Engineer in Lahore
AI/ML ENGINEER
From models to production.

I build AI systems that connect models, data, APIs and backend infrastructure into products that work beyond the prototype stage.

Portfolio
EXPLORE WORK
AI & Backend Core Focus
Zeploy Tech Founder
2 Industry certifications
SVYPE AI/ML Engineer
About

Engineering at the edge of AI

I build AI systems designed to work beyond the prototype stage.

My work sits at the intersection of AI/ML engineering, LLM integration and backend architecture: connecting models, data pipelines, APIs and software infrastructure into products that serve real workflows.

At SVYPE, I work as an AI/ML Engineer, bringing machine learning and LLM capabilities into internal platforms and client products. I joined as a Full Stack Development Intern and moved into the AI/ML role in 2026.

As Founder & Lead Engineer of Zeploy Tech, I lead the engineering and delivery of web, backend and AI products, including zeploy.tech and its AI assistant, Zee.

My final-year project, MockAI, is a multimodal interview-evaluation system that brings NLP, speech recognition and facial analysis into one evaluation pipeline, with separate candidate and admin platforms.

I treat AI as a systems engineering problem: a good model is only the start; latency, reliability and architecture decide whether it works in production.

EducationBS Computer Science, COMSATS University Islamabad, 2023 to 2027
LocationLahore, Pakistan
RoleAI/ML Engineer at SVYPE · Founder & Lead Engineer at Zeploy Tech
FYPMockAI: multimodal AI interview evaluation (completed)
LanguagesEnglish (Fluent) · Urdu (Native)
CourseworkMachine Learning · Computer Vision · NLP · Database Systems
Skills

The tools behind the systems

Grouped by layer, from the model to the API to the interface.

AI Systems
Machine Learning Deep Learning NLP LLM Integration Gemini API DistilBERT Speech-to-Text FERPlus scikit-learn Pandas NumPy
Backend
FastAPI Node.js Express REST API Design JWT Auth Redis Background Jobs
Frontend
React.js Next.js TypeScript Tailwind CSS Three.js GSAP Flutter
Databases
PostgreSQL MongoDB Prisma
Tools
Git GitHub GitHub Actions (CI) Docker FFmpeg Postman Linux VS Code
Languages
Python TypeScript JavaScript SQL Dart Java C#
Projects

Systems I've built

AI systems, backend platforms and product engineering work.

MockAI multimodal AI interview evaluation platform by Syed Asjad Abbas
01 / AI System Completed

MockAI

ChallengeEvaluate interview performance across answer content, speech and facial expression at the same time, without manual review.
SolutionA multimodal pipeline that scores each recorded answer on content (DistilBERT semantic scoring), delivery (Google Speech-to-Text with pace, pause and filler analysis) and composure (YuNet face detection with FERPlus emotion analysis). A late-fusion engine weights the three 50 / 30 / 20 into an explainable report. Built with separate candidate and admin platforms on a FastAPI and MongoDB backend.
React.js → FastAPI → FFmpeg → DistilBERT · Speech-to-Text · FERPlus → Late fusion
Python FastAPI React.js MongoDB DistilBERT FERPlus FFmpeg
ZePlay adaptive video streaming platform by Syed Asjad Abbas
02 / Video Streaming Platform Internship Project

ZePlay

ChallengeDeliver video reliably across different network conditions without blocking the application while videos are processed.
SolutionUploads are transcoded in background FFmpeg jobs into 1080p, 720p and 480p HLS renditions for adaptive-bitrate playback. A FastAPI backend handles authentication, profiles, watch history, search and recommendations, with Redis caching in front of PostgreSQL. Load-tested with k6 at 650 requests per second with 500 concurrent users and 0% failures (local benchmark, 4 workers); 59 of 59 backend tests pass.
React + TS → FastAPI → HLS.js / FFmpeg → Redis → PostgreSQL
React TypeScript FastAPI PostgreSQL Redis HLS.js FFmpeg Docker
FlightOne travel platform and AI assistant by Syed Asjad Abbas
03 / AI Travel Platform Full-Stack Architecture

FlightOne

ChallengeMake AI part of the travel workflow, from a traveller's request to booking and post-booking support, instead of a chat window attached to a website.
SolutionAva, an AI travel consultant, turns a request like "Lahore to Karachi on 10 Oct" into comparable options with filters and recommendations, then leads into a structured checkout (traveller details, payment, ticket). After booking, journey tracking, disruption assistance and servicing continue in the same product.
Next.js Web → React Client → Express API → AI Agent → PostgreSQL
Next.js React TypeScript Express Prisma PostgreSQL AI Assistant
Serene Management hotel PMS dashboard and mobile housekeeping board by Syed Asjad Abbas
04 / Hospitality Platform Built & Tested

Serene Management

ChallengeRun daily hotel operations for one or several properties without staff seeing another property's data or billing records ending up inconsistent.
SolutionA hotel property management system covering reservations, guest profiles, rooms, housekeeping, rates, billing, reporting and night audit. Permissions are enforced on the server, each property's data is isolated, billing and night audit run inside database transactions, and every operational change is recorded in an audit trail. Payments use idempotency keys and night audit runs in a background worker. Performance fixes took global search from 8.3 s to 0.45 s and sustained load from 10–25 to 50 requests per second with 0 errors (P95 192 ms, single instance).
Next.js → Server-side permissions → Prisma → PostgreSQL
Next.js React TypeScript Prisma GitHub Actions CI
Zee, the Zeploy Tech AI assistant, promotional showcase
05 / AI Assistant Live

Zee (Zeploy Tech)

ChallengeGive website visitors instant, reliable answers about Zeploy's services without the assistant ever failing visibly.
SolutionAn AI assistant on zeploy.tech that streams responses from Google Gemini with a focused system prompt, cleans conversation history before each request, and automatically falls back to a second model if the primary one is unavailable.
React → Server route → Gemini API (streaming) → Fallback model
TypeScript React Gemini API Streaming LLM Integration
Archive & Client Projects

Additional Work

Selected web applications, client websites and earlier projects.

3D Web Experience · LIVE

Serene Heights 3D

Scroll-driven cinematic web experience merging interactive WebGL/Three.js 3D models, custom shader atmospheres, GSAP animation timelines, and smooth Lenis scrolling.

React TypeScript Three.js WebGL GSAP Lenis Vite
Mobile App · Internship

TuneBox

Flutter music player with playback controls, playlist management and Provider state management.

Flutter Dart Provider REST APIs
Academic Project

EduCart

MERN e-commerce platform with JWT authentication, PayPal checkout and an admin dashboard.

React.js Node.js Express MongoDB JWT Tailwind CSS
Academic Project

Gym Management System

Desktop gym management system with member profiles, payments, measurements and a store module.

C# Windows Forms SQL Server
Commercial Website

FrameGrill

Polished restaurant website focused on premium presentation, responsive interaction, and commercial web execution across desktop and mobile.

HTML5 CSS3 JavaScript jQuery Swiper JS
Corporate Platform

Serene Vantage

Next-generation corporate platform for a diversified development group operating across Technology, Energy, and Property with editorial aesthetics and choreographed motion.

Next.js 15 React 19 TypeScript Tailwind CSS Framer Motion
Commercial Website

Italian Construction Website

Premium construction company website focused on modern architecture, responsive layouts, visual presentation, and business conversion.

HTML5 CSS3 JavaScript UI Design
Experience

Where I've worked

Jul 2026 — Present
SVYPE
Lahore, Pakistan (Hybrid)
AI/ML Engineer
Previously Full Stack Development Intern (May 2026 to Jul 2026)
  • Develop and integrate AI capabilities across internal platforms and client-facing products, covering model integration, data preparation, evaluation and backend implementation.
  • Integrate LLMs and machine learning models into existing products through Python backend services and APIs.
  • Joined as a Full Stack Development Intern building JavaScript web applications and REST APIs, and was promoted to AI/ML Engineer on completion of the internship.
Apr 2026 — Present
Lahore, Pakistan (Remote)
Founder & Lead Engineer
  • Founded and lead a software studio delivering web, backend and AI products for clients.
  • Own the technical lifecycle from requirements and system architecture through implementation, integration and deployment.
  • Designed and built zeploy.tech and Zee, its Gemini-powered AI assistant.
Jan 2021 — Present
100 Se Support
Khairpur District, Sindh (Hybrid)
Co-Founder & COO
  • Co-founded a community welfare initiative that connects people in need with verified support through a transparent, accountable process.
  • Lead day-to-day operations: processes for receiving and verifying support requests, team and volunteer coordination, and accountable records.
Certifications

Credentials

Google AI Essentials

Google & Coursera
Professional Certificate 2026

Deep Learning Specialization

DeepLearning.AI & Coursera
Specialization Certificate 2026
At a Glance

By the Numbers

3
AI models in the MockAI evaluation pipeline (NLP · Speech · Vision)
7.7×
ZePlay throughput gain under load (84 → 650 req/s, k6 benchmark)
2
Industry certifications (Google · DeepLearning.AI)
2
Companies founded or co-founded (Zeploy Tech · 100 Se Support)

Currently Building

What's Next

I'm building Intelligence OS, a multi-tenant RAG platform with hybrid search, reranking and a built-in evaluation framework. It will be published as a full case study on completion.

Contact

Let's build something real.

Open to AI engineering roles, technical partnerships and client projects through Zeploy Tech.

asjadabbaszaidi@gmail.com