Skip to content
Talha ZainApplied AI Engineer

Applied AI Engineer · Islamabad, Pakistan

Mode / Production

I architect AI systemsthat reason, retrieve,orchestrate & act.

Production-grade agentic AI, LLM, RAG, multimodal and machine-learning systems—from model orchestration to APIs, evaluation and deployment.

Explore Selected Systems
Let's TalkGitHub

Agentic AI · LLM Systems · Retrieval-Augmented Generation · Production AI Engineering

Talha Zain — Applied AI Engineer

TZ Systems Graph

Conceptual production topology

Live
ORCHESTRATORLangGraph

Stateful routing across agents, retrieval, tools and models.

Hover · tap · arrow keys — paths illuminate with the active node

Production signals

Measured outcomes from shipped systems.

Figures tied to specific systems—not vanity company KPIs.

30+
Image Models
20+
Video Models
5+
Audio Models
290+
UAE POI Categories
< 5s
Recommendation Response
3–4 hrs/week
Manual Work Eliminated

Selected systems

Production systems across generative AI, healthcare, retrieval and data infrastructure.

Flagship case studies. Additional systems live in the Engineering Archive.

01Generative AIproduction

AI Compare Hub

Production multimodal generation platform giving users unified access to a large ecosystem of image, video, and audio models—with side-by-side comparison from a single prompt.

30+ Image models20+ Video models5+ Audio models
AI Compare Hub recent generations dashboard with multimodal model outputs and asset history.

Unified multi-model generative media platform.

Case study
02Agentic AIinternal

MedicAI

LangGraph-orchestrated clinical platform with specialized agents, Agentic RAG, tenant isolation, human-in-the-loop gates, and evaluation via LangSmith and RAGAS.

Multi-agent LangGraph orchestrationHITL Human approval gatesRBAC Tenant isolation
MedicAI clinical operations dashboard with AI agents, appointments, and hospital analytics.

Multi-tenant multi-agent clinical AI platform.

Case study
03Intelligent Searchproduction

MediaX

LLM-driven system that transforms natural-language campaign requirements into structured geographic intent, retrieves inventory, and ranks placements across 290+ UAE categories in under five seconds.

290+ UAE POI categories< 5s Recommendation response30% Footfall weight
MediaX map interface for finding optimal outdoor media spaces across the UAE.

Semantic and geospatial OOH recommendation engine.

Case study
04Data Systems / ML Systemsinternal

ECG Intelligence Platform

Unified heterogeneous PDF, image, CSV, and waveform inputs into a canonical clinical schema with configurable field mapping—eliminating recurring manual analyst work and enabling analytics/ML pipelines.

3–4 hrs/week Manual work eliminated4 Input modalities
ECG Data Unification enterprise architecture spanning context engine, platform, and delivery skills.

Vendor-agnostic clinical data unification for ML readiness.

Case study
View all systems

Engineering capabilities

Specialization first. Supporting stack second.

Agentic AI, LLM systems, RAG, and production engineering define the practice. ML, cloud, and data infrastructure support delivery.

01 · Primary

Agentic Systems

Orchestrated multi-agent workflows with state, tools, and human gates.

LangGraphLangChainMulti-Agent SystemsAgentic RAGTool CallingHuman-in-the-LoopStructured Outputs
02 · Primary

LLM Infrastructure

Model access, retrieval, evaluation, and observability for production LLM systems.

OpenAIClaudeLlamaHugging FaceEmbeddingsRetrievalRAGASLangSmithObservability
03

Machine Learning

Classical and deep learning for prediction, language, and vision workloads.

PyTorchTensorFlowScikit-learnXGBoostLightGBMNLPComputer VisionForecastingClassificationRegressionClustering
04 · Primary

Production Engineering

APIs, data stores, async systems, and cloud deployment for AI products.

PythonFastAPIRESTAsync SystemsPostgreSQLpgvectorRedisMongoDBDockerCI/CDAWSGCPAzureETL
Capability graph / production AI
Center
PRODUCTION AI

Agentic Systems

LangGraphMulti-AgentTool CallingHITL

Engineering trajectory

Ownership compounding over time.

From foundational NLP services to production agentic and generative systems.

  1. 01

    BoolMind

    Jul 2025 — PresentPresent

    ML Engineer — Applied / Generative AI

    Architected an LLM-powered semantic advertising recommendation engine translating campaign requirements into structured geo-intent and ranking OOH inventory across 290+ UAE categories.

  2. 02

    IT Genics

    Jan 2024 — Jun 2025

    Associate AI Engineer

    Built no-code conversational and voice AI workflows integrating Llama, NLP pipelines, Twilio, and business actions—packaged behind production APIs and containerized services.

  3. 03

    IT Solutions World Wide

    Jul 2023 — Dec 2023

    Junior AI Engineer

    Built NLP applications with GPT-3, Llama, BERT, and Falcon, applying transfer learning and fine-tuning to move experiments into reusable services.

Full trajectory

How I build

From model capability to production system.

Methodology over tool lists—understand, design, orchestrate, evaluate, ship, observe.

  1. 01

    Understand

    Business problem / constraints

  2. 02

    Design

    Architecture / data / tools

  3. 03

    Orchestrate

    Models / agents / retrieval

  4. 04

    Evaluate

    Quality / latency / reliability

  5. 05

    Ship

    APIs / containers / cloud

  6. 06

    Observe

    Tracing / monitoring / iteration

About

AI engineering is not about calling a model.

I focus on the systems surrounding the model—retrieval, orchestration, state, tools, data, evaluation, reliability, APIs and deployment—because that is what turns an AI capability into a production product.

Islamabad, Pakistan · Applied AI Engineer

Next collaboration

Building something
AI-native?

Interested in complex problems involving agents, LLM systems, RAG, multimodal AI and production ML.

Start a ConversationEmailLinkedInGitHub