Software engineer · AI systemsOpen worldwide

I buildAI systemsthat ship.

I build production-grade AI and full-stack systems. Lately: benchmark-driven agentic evaluation, RAG pipelines, and high-throughput backends — and I’m as happy debugging a Django query as I am fine-tuning an LLM.

Explore the workRésumé
LLM evaluationRAG systemsFull-stack productsCloud infrastructureProduction reliability
01Selected work

From model behavior to production throughput.

Five projects across conversational AI, applied machine learning, backend modernization, content automation, and location analytics.
01 / 05AI Companion for Calm & Clarity

SereniMind

An empathetic AI chatbot with deep conversational memory. React + FastAPI, with Deepseek-R1 via Ollama, secure auth and history storage. Cut hallucinations 25% through contextual memory.

ReactFastAPIDeepseek-R1Ollama
25%
hallucination reduction
Selected projectView repository
02 / 05AI for Connectivity · Lablab.ai Finalist

Enhancing Infrastructure & Accessibility

Trained XGBoost, Random Forest and K-Means on John Snow Labs and World Bank datasets. Integrated Mistral-7B-Instruct-v0.2 for recommendations in a React + Flask full-stack experience.

XGBoostMistral-7BReactFlask
3
model families trained
Selected projectView repository
03 / 05Legacy → Django Microservices

pyERP

Migrated a monolith to Django microservices with ETL pipelines that cut redundancy by 30%. Built Vue.js modules that reduced critical errors by 45% in the first cycle.

DjangoVueMicroservicesETL
45%
fewer critical errors
Mar 2025 – Jun 2025
ArchiveAdditional builds
04
AI SEO Content Engine

BotBlogR

Re-architected a Flask app to FastAPI + React for a 40% latency drop. Integrated the OpenAI API for SEO content generation — 70% faster, with PostgreSQL queries tuned for +30% RPS.

FastAPIReactOpenAIPostgreSQL
70%
faster content generation
Nov 2023 – Feb 2024
05
Crime-Tracking Platform

IntelPort

Location-based analytics with Django + React + PostgreSQL. Engineered REST APIs to cut API latency by 30%, and shipped secure auth with 99.9% uptime.

DjangoReactPostgreSQLREST
99.9%
secure auth uptime
Sep 2023 – Dec 2023
github.com/MateehUllahBrowse all repositories
02Operating profile

One engineer across
model, API and interface.

The best AI products are not model demos. They are complete systems where data, backend performance, interface design and reliability all work together.

Software engineer with 3+ years of experience building production LLM systems and scalable full-stack applications. Engineered SFT, RLHF, benchmark-driven agentic evaluation, and model deployment — cut hallucinations by 7%, improved workflow efficiency by 20%, and validated AI connector integrations at 99.9% uptime.

More about how I work
Base
Austria
Reach
Remote · worldwide
Focus
AI systems · product engineering
Status
Open to opportunities worldwide
01

Model quality

AI systems

Evaluation, SFT, RLHF and retrieval workflows designed around measurable behavior in production.

02

System design

Backend engineering

Fast, reliable APIs and data pathways built with Python, Django, FastAPI, PostgreSQL and Redis.

03

Product layer

Full-stack delivery

Interfaces that turn complex systems into clear products across React, Next.js, Vue and TypeScript.

04

Operations

Cloud delivery

Reliable deployments and automation across AWS, GCP, Docker, Kubernetes and CI/CD workflows.

03Service record

Engineering outcomes
across AI and product.

Work spanning analytics APIs, benchmark-driven model evaluation, SaaS connector reliability, and full-stack platforms.
01

Mar 2025 – Feb 2026

Software Engineer

Footmetrics

Designed and shipped data pipelines and APIs powering customer analytics, with a focus on latency, throughput, and operational efficiency.

  • Orchestrated 12+ Django REST endpoints to ingest, process and deliver customer analytics.
  • Forged backend pathways to process and serve customer data, reducing operational latency by 40%.
  • Optimized DB operations via strategic caching, reducing server load by 30% and improving query efficiency.
DjangoPythonPostgreSQLRedisREST APIs
02

Sep 2025 – Feb 2026

LLM Engineer

TuringRemote · United States

Built benchmark-driven workflows and evaluation pipelines for agentic AI systems in the fund management vertical.

  • Spearheaded a benchmark-driven workflow for agentic AI systems — reduced hallucinations by 7% and lifted task success across fund management.
  • Designed scenarios for fund management, investor onboarding and audit-trail workflows, achieving 20% efficiency gains.
  • Deployed LLMs for real-time risk assessment — cut false positives by 10% with higher overall accuracy.
  • Authored and reviewed multi-language code (Python, JS, SQL) for LLM training datasets.
LLMsSFTRLHFPythonEvaluation
03

Sep 2025 – Jan 2026

AI Engineer

Micro1Remote · United States

Validated and stress-tested AI connector integrations across major SaaS platforms, automating QA into CI/CD.

  • Validated AI connector integrations for Zoho CRM, Freshdesk and Salesforce at 99.9% uptime.
  • Populated synthetic business data simulating real-world conditions, improving test coverage by 30%.
  • Automated connector testing with Python scripts integrated into CI/CD — saved 10+ hours per week.
PythonZohoSalesforceFreshdeskCI/CD
04

Jun 2023 – Mar 2025

Associate → Software Engineer

Devsinc

Built internal-tool frontends, reusable component systems and customer engagement integrations supporting 1000+ users.

  • Constructed React.js interfaces for internal workflows integrated with Django services — supported 1000+ internal users.
  • Programmed Twilio APIs for SMS, WhatsApp & calls — elevated response rates by 70%.
  • Modernized user-facing features and component design, reducing page load times by 40%.
  • Shipped Vue.js + TypeScript features and engineered 20+ reusable components — cut reported frontend bugs by 15%.
  • Automated 100+ regression tests with Robot Framework and built scalable Hypermea + Docker services — reduced deployment time by 30%.
ReactVueTypeScriptDjangoTwilioDocker
04The workbench

Tools chosen for
production trade-offs.

A pragmatic stack across model quality, backend systems, product interfaces, data, cloud, and delivery automation.
01

Languages

PythonTypeScriptJavaScriptSQLC/C++Shell
02

AI / ML

LLMsSFTRLHFHugging FaceScikit-learnXGBoostK-Means
03

Backend

DjangoFastAPINestJSHypermeaREST APIs
04

Frontend

ReactNext.jsVueTailwind CSS
05

Databases

PostgreSQLMongoDBMySQLSQLite
06

Cloud / DevOps

AWSGCPDockerKubernetesLambdaS3EC2
07

Testing & Tools

Robot FrameworkBrowser LibraryGitCursorClaude CodeGitHub Copilot
Model layerAPI layerProduct layerCloud layer
05 / Credentials

The foundation behind the work.

Education

Bachelor of Science in Computer Science

FAST-NUCES · Lahore, Pakistan · 2019 – 2023

Recognition

  • Dean’s List of Honors — FAST-NUCES (Jun 2023)
  • Finalist — AI for Connectivity Hackathon (Lablab.ai)
06 / Start a conversation Austria · available worldwide

Let's buildbeyondthe demo.

I'm open to opportunities across AI systems, backend engineering, and full-stack product development.

Start a conversationConnect on LinkedIn
Emailmateehullah89@gmail.comLinkedInmateeh-ullahGitHubMateehUllahUpworkFreelance profile