Asfand Yar.
Resume

SOFTWARE · BACKEND · AI

Asfand Yar

I build backend services, APIs, and AI-powered applications — connecting LLMs, RAG pipelines, and agents with the software that makes them useful.

View projects View resume
GitHub LinkedIn Get in touch
THE ANATOMY OF A SYSTEMFIG. 01
Data→ Intelligence→ Application

Open to opportunities / Islamabad, Pakistan

EXPLORE THE PORTFOLIO
BACKEND SYSTEMS+GENERATIVE AI+RAG & AGENTS+PRODUCTION SOFTWARE

01 / ABOUT

Built on engineering. Driven by possibility.

BACKGROUND
Computer ScienceBUITEMS · Expected Dec 2026
FOCUS
Backend & generative AIAPIs, retrieval, and agents
BASED IN
Islamabad, Pakistan

My work sits at the intersection of backend engineering and generative AI.

I build Python and FastAPI backends, integrate language models, and develop retrieval pipelines that give AI applications the context they need. My experience spans enterprise AI features, model evaluation, data processing, and client web applications.

I’m studying Computer Science at BUITEMS, with graduation expected in December 2026. I’m interested in the complete path from data to a useful product: how an API behaves, where context comes from, and how the pieces work together in production.

Reliable backendsContext-aware AIPractical software

02 / EXPERIENCE

Applied in the real world.

Industry experience across AI backends, model evaluation, and web development.

Jan 2026 – Mar 2026

Emumba

Islamabad, Pakistan · On-site

Gen AI Backend Developer

MOST RECENT
  • Developed backend systems using Python and applied advanced prompt engineering to optimize Large Language Model outputs for enterprise applications.
  • Engineered Retrieval-Augmented Generation (RAG) pipelines to enhance model accuracy and contextual query answering.
  • Collaborated with cross-functional teams to deploy scalable AI features, ensuring low-latency data processing and API integration.

Python · Generative AI · LLMs · RAG · AI Agents · Prompt Engineering · FastAPI · REST APIs · Backend Development · Docker · API Integration

Mar 2025 – May 2025

Developers Hub Corporation

Remote

AI Intern

  • Developed Python scripts to extract, clean, and format raw data, creating high-quality datasets for testing internal machine learning models.
  • Conducted foundational prompt engineering and benchmarked various open-source language models to document performance and response accuracy.
  • Built internal UI dashboards and automated scripts to visualize AI outputs and test API endpoints, streamlining day-to-day development tasks.

Python · Prompt Engineering · Data Processing · Open-source LLMs · API Testing

Jul 2025 – Aug 2025

CodeAlpha

Remote

Frontend Development Intern

  • Developed interactive frontend features and landing pages, focusing on mobile-first design principles.
  • Translated design mockups into clean, semantic HTML/CSS and JavaScript code.
  • Assisted in debugging UI issues and improving website performance metrics.

HTML · CSS · JavaScript · Mobile-first · UI Debugging · Frontend Architecture

03 / SELECTED WORK

Ideas, engineered into systems.

A closer look at the applications, pipelines, and tools I’ve built.

FINAL YEAR PROJECT · AI & FULL-STACK ENGINEERING

Intervexa

Practice the interview. Understand the feedback.

An AI-powered mock interview platform combining adaptive questions, multimodal feedback, and live interviews with human interviewers. Built to bring interview practice and evaluation into one workflow.

Next.jsTypeScriptExpress.jsMongoDBPython / FlaskWebRTCSocket.IO

INTERVIEW → ANALYSIS → FEEDBACK

  1. 01 / PRACTICE

    AI or human-led interview

    Adaptive questions · Live video sessions

  2. 02 / ANALYZE

    Text, voice & visual signals

    NLP · Speech processing · Computer vision

  3. 03 / REVIEW

    Feedback in one place

    AI reports · Human evaluation

NEXT.JS ↔ EXPRESS / MONGODB ↔ PYTHON AI

Explore the technical approach
Interview practice
Role- and experience-based question generation through Groq / LLaMA and Gemini, alongside human-led interview sessions.
Multimodal evaluation
A Python AI gateway combines answer analysis, Whisper transcription, voice analysis, and facial-expression processing using tools including Sentence-Transformers, Librosa, DeepFace, and OpenCV.
Live sessions
WebRTC handles video, audio, and screen sharing, with Socket.IO signaling and an Express / MongoDB backend for the interview workflow.
CONTEXT-AWARE CONVERSATION01
SPEECHINPUTCONTEXTRETRIEVALLOCAL LLMLM STUDIOREPLYVECTOR STORE
Speech → retrieval → generationCONCEPTUAL FLOW

01 / VOICE · RETRIEVAL · LOCAL AI

AI Powered Voice Agent

Context-aware conversation without a cloud dependency. A voice agent that retrieves relevant information and generates responses with a local language model.

PythonLM StudioRAGLangChainVector DBSpeech API
Explore the technical approach
What I built
A conversational voice agent that connects speech processing, retrieval, and local LLM generation.
Technical approach
Conversational voice agent built with RAG and a local LLM served via LM Studio. Processes user speech, retrieves context from a vector store, and generates accurate responses — fully offline, zero cloud dependency.
Retrieval & model
LangChain, a vector store, and a local LLM served through LM Studio.
Engineering focus
Connecting retrieved context with speech-driven queries while keeping the application offline-capable.
PROMPT EVALUATION WORKFLOW02
INPUTPROMPTFEW-SHOTTEMPLATESMODELEVALUATIONRESULTPERFORMANCE TRACKING
Templates → evaluation → iterationCONCEPTUAL FLOW

02 / LLMS · EVALUATION · BACKEND

LLM Prompt Optimization Engine

A framework for making prompt iteration measurable, with automated evaluations and performance tracking across model versions.

PythonOpenAI APIPrompt EngineeringFastAPIPostgreSQL
Explore the technical approach
What I built
A prompt optimization framework with automated evaluation pipelines and few-shot template generation.
Technical approach
Prompt engineering framework to benchmark and optimize LLM outputs. Includes automated evaluation pipelines, few-shot template generation, and performance tracking across model versions.
Backend & data
Python, FastAPI, the OpenAI API, and PostgreSQL.
Engineering focus
Comparing prompt strategies and tracking language model performance across versions.

CLIENT APPLICATIONS

Healing Hands of Virginia

Client website for a healthcare practice in Virginia. Clean, accessible design with appointment booking, service listings, and mobile-first layout.

Web Development · Responsive Design · SEO · UI/UX

my1ai.app

AI-powered web application for clients. Features intelligent automation, a clean user interface, and seamless API integrations for real-world AI use cases.

AI Integration · Web App · API · UI/UX

More things I’ve built 05 PROJECTS

AI Dataset Pipeline

Automated ETL pipeline to extract, clean, and format raw datasets for machine learning model training and benchmarking.

Python · Pandas · NumPy · REST API · Data Processing

Python AI Chatbot

Intelligent chatbot with natural language processing using NLTK and intent classification, backed by a Flask REST API.

Python · NLTK · TensorFlow · Flask · REST API

Developer Portfolio

This portfolio: a statically rendered Next.js website with accessible navigation, light and dark themes, and a lightweight interactive computational visual.

Next.js · TypeScript · CSS · SVG

Snake Game Arcade

Classic Snake game built in Python with Pygame. Features smooth controls, score tracking, increasing speed levels, collision detection, and a retro arcade-style UI.

Python · Pygame · Game Dev · OOP

URL Changer React App

Dynamic React application for URL manipulation and redirection with real-time preview, input validation, copy-to-clipboard, and a clean responsive UI.

React · JavaScript · React Router · Tailwind CSS

Find me on GitHub

04 / THE STACK

The right tools for the system.

An engineering toolkit shaped by hands-on work, from the database to the interface.

01

Backend engineering

Services, APIs & application logic

  • Python
  • FastAPI
  • REST APIs
  • Node.js
  • Authentication
  • WebSockets
02

AI & generative systems

Retrieval, models & intelligent workflows

  • LLM Integration
  • RAG Pipelines
  • Prompt Engineering
  • OpenAI
  • AI Agents
  • LangChain
  • Data Processing
  • LM Studio
  • n8n
03

Databases & storage

Relational, document & vector data

  • Redis
  • PostgreSQL
  • MySQL
  • MongoDB
  • Vector DBs
  • Firebase
04

Frontend

The interface to the system

  • React.js
  • Next.js
  • TypeScript
  • JavaScript ES6+
  • Tailwind CSS
  • HTML & CSS
  • Framer Motion
05

Infrastructure & tools

Building, testing & shipping

  • Git & GitHub
  • Docker
  • Claude Code
  • Codex
  • Antigravity
  • Postman
  • VS Code
  • Figma
  • Jupyter Notebook

05 / HOW I CAN HELP

From a backend to an AI workflow.

01 / SOFTWARE

Backend & web development

Python and FastAPI services, API integrations, and responsive client applications built with React and Next.js.

02 / INTELLIGENCE

AI systems & integrations

RAG pipelines, AI agents, voice interfaces, prompt engineering, and local or cloud LLM integration.

03 / WORKFLOWS

Data & automation

ETL and data processing, model benchmarking, n8n workflows, and Docker-based application deployment.

06 / CONTACT

Have something
worth building?

I’m open to software engineering roles and freelance projects. Get in touch to discuss the work.

asfandyar273263@gmail.com

Based in Islamabad, Pakistan

START A CONVERSATION Email LinkedIn GitHub WhatsApp Download resume