ABDUL WASAY.

Generative AI Engineer

LLMs • Autonomous Agents • Enterprise RAG

LET'S TALK ↗

LARGE LANGUAGE MODELS

MULTI-AGENT SYSTEMS

ENTERPRISE RAG

ADVERSARIAL RED-TEAMING

LANGCHAIN & FASTAPI

CEDAR WASM GATEWAYS

VECTOR DATABASES

DOCKER & AZURE SCALE

LARGE LANGUAGE MODELS

MULTI-AGENT SYSTEMS

ENTERPRISE RAG

ADVERSARIAL RED-TEAMING

LANGCHAIN & FASTAPI

CEDAR WASM GATEWAYS

VECTOR DATABASES

DOCKER & AZURE SCALE

WHAT I DO /

(Services)

I architect production-grade Generative AI applications, multi-agent reasoning workflows, and resilient adversarial defense pipelines that scale effortlessly from research to enterprise infrastructure.

(01)

Generative AI & Multi-Agent Systems

From autonomous task routing to multi-step reasoning pipelines. I engineer intelligent agent systems capable of automated triage, tool calling, chain-of-thought analysis, and complex operational decision-making.

01 Multi-Agent Orchestration & LangChain

02 Autonomous Tool Calling & Reasoning Loops

03 Automated IT Triage & Root-Cause Analysis

(02)

Enterprise RAG & Vector Intelligence

Precision retrieval architectures for massive internal enterprise knowledge. Combining hybrid vector search, BM25 dense-sparse matching, and semantic reranking to eliminate hallucinations and accelerate document Q&A.

01 Naive, Hybrid & Advanced RAG Pipelines

02 ChromaDB, FAISS & High-Throughput Embeddings

03 Context Compression & Re-ranking Optimization

(03)

AI Safety, Governance & Cloud Microservices

Hardening enterprise LLM systems against adversarial threats. Implementing inline policy gateways (Cedar WASM), multi-strategy jailbreak neutralizers, and production microservices with Docker on Microsoft Azure.

01 MCP Governance Gateways & Cedar WASM Policies

02 Adversarial Jailbreak Detection & Red-Teaming

03 FastAPI, Docker Containers & Azure Cloud Deployment

03 // FLAGSHIP ARCHIVES

SELECTED WORKS /

Production-grade AI safety gateways, knowledge graph reasoning agents, and live cloud applications. Explore right-to-left.

01 / 03

sentinel_policy.cedar — WASM Sandbox

Cedar WASM Engine

permit (
principal in Role::"autonomous_agent",
action in [Action::"execute_tool", Action::"query_vector"],
resource in Resource::"production_db"
) when {
context.threat_score <= 0.05 && context.audit_chained == true
};

                      [VERIFIED] 21/21 Agentic Threat Vectors Neutralized • SHA-256 Audit Chained
                    

MCP Sentinel — AI Agent Security Gateway

2026

Inline Model Context Protocol (MCP) governance gateway engineered to mitigate 21 threat vectors, including tool poisoning, unauthorized lateral movement, and indirect prompt injection. Features the Cedar policy engine compiled to WebAssembly (WASM), real-time human-in-the-loop (HITL) approvals, SHA-256 chained tamper-evident audit trails, and an SRE observability dashboard backed by 620+ automated tests.

                      Core Engine
                      Cedar Policy (WASM)
                    

                      Threat Scope
                      21 Attack Vectors
                    

                      Audit Proof
                      SHA-256 Hash Chain
                    

                      Verification
                      620+ Passing Tests
                    

AI Agent Security MCP Gateway Cedar WASM Python 3.11 TypeScript Docker SRE Observability

View Repository ↗ Architecture Specs ↗

career_graph_rag.py — Graph Query Stream

Knowledge Graph Engine

# GraphRAG Entity Resolution & Trajectory Mapping
async def resolve_career_path(target_role: str, user_skills: list):
graph_nodes = await knowledge_graph.extract_subgraph(target_role)
skill_distance = vector_index.compute_semantic_gap(user_skills, graph_nodes)
roadmap = await gemini_client.generate_milestones(skill_distance, depth=4)
return GraphResponse(trajectory=roadmap, topology_nodes=graph_nodes.count)

                      [RESOLVED] 1,420 Career Nodes • 4,890 Vector Edges • Zero Hallucinations via Grounded Graph
                    

Smart Career Advisor — GraphRAG Agent

2026

Enterprise-grade Career Advisory Agent utilizing GraphRAG (Knowledge Graph entity mapping + vector embeddings) integrated with Google Gemini 2.0. Constructs interactive 3D career trajectory graphs, performs semantic skill gap analysis against enterprise role benchmarks, and synthesizes deterministic milestone roadmaps with local embeddings.

                      Architecture
                      GraphRAG + Vector DB
                    

                      LLM Backbone
                      Google Gemini 2.0
                    

                      Visualization
                      3D Trajectory Graphs
                    

                      Embeddings
                      Local Vector Search
                    

GraphRAG Google Gemini 2.0 Knowledge Graphs FastAPI 3D Data Topology Python Network Analysis

View Repository ↗ Interactive Specs ↗

https://ai-learning-app-two.vercel.app

⚡ Production Live

AI LEARNING STUDIO

8 Active Courses Adaptive Quizzes Persistent Chat Memory

Module 04 // LLM Architecture

Transformers, Self-Attention & Tokenization

AI Tutor [Groq 750 T/s]

"In multi-head attention, queries and keys project onto orthogonal subspaces before softmax score calculation..."

AI Learning Studio — Ed-Tech Platform

2025

Full-stack production educational platform engineered to teach modern Artificial Intelligence. Features 8 comprehensive courses, an intelligent conversational AI tutor with persistent memory, real-time adaptive quizzes, and student mastery metrics. Shipped with Next.js, Firebase Auth & Firestore, and Groq ultra-low-latency LLM inference.

                      Live Deployment
                      ai-learning-app-two.vercel.app
                    

                      Course Library
                      8 Comprehensive Modules
                    

                      AI Tutor Core
                      Groq LLM + Persistent Memory
                    

                      Full Stack
                      Next.js 14 • Firebase • Tailwind
                    

Next.js 14 React Firebase Auth Firestore Groq API TailwindCSS Production SaaS

Launch Live Platform ↗ View Repository ↗

04 // THE ARCHITECT & ARSENAL

PHILOSOPHY & SKILLS /

Bridging probabilistic foundation models with deterministic software architecture, adversarial robustness, and production resilience.

// GENERATIVE AI ENGINEER

ABDUL WASAY

I engineer production-grade LLM systems, scalable autonomous agents, and adversarial safety layers that transform frontier AI research into battle-tested software.

My engineering philosophy focuses on reliability, verifiability, and low-latency execution. Rather than relying on simple prompt wraps, I construct hardened architectures incorporating WebAssembly policy gates, deterministic knowledge graphs, and hybrid lexical-semantic retrieval engines.

Currently pursuing a BS in Software Engineering at Iqra University (2023–2027) while deploying mission-critical systems across governmental aviation data infrastructure and production AI startups.

Certified by IBM on Coursera with the IBM Generative AI Engineering Professional Certificate (July 2026).

• Status: Available for GenAI Roles

Base: Karachi, Pakistan (UTC+5)

Credentials: IBM Certified • 620+ Unit Tests

Foundation Models & Agentic AI

Inference & Reasoning

Google Gemini 2.0 Claude 3.5 Qwen 2.5 LangChain Autonomous Multi-Agent Swarms Chain-of-Thought (CoT) Prompt Engineering & Structured Outputs Groq High-Throughput Inference

Adversarial Safety & Red-Teaming

AI Defense

Model Context Protocol (MCP) Cedar Policy Engine (WASM) Jailbreak Detection & Neutralization Prompt Injection Defense Adversarial Red-Teaming Suites SHA-256 Chained Audit Trails Human-in-the-Loop (HITL) Gates

Vector Search & Information Retrieval

Zero Hallucinations

GraphRAG & Knowledge Graphs ChromaDB FAISS Vector Search BM25 Sparse Lexical Search Cross-Encoder Re-Ranking Reciprocal Rank Fusion (RRF) Semantic Chunking & Embeddings

Full-Stack & Cloud Infrastructure

Production Systems

Python (Advanced 3.11+) FastAPI Microservices TypeScript & JavaScript Next.js 14 / React ASP.NET (C#) SQL Server & Relational Schema Docker Containerization Microsoft Azure Git & CI/CD Pipelines

05 // TRAJECTORY & CREDENTIALS

EXPERIENCE /

Delivering mission-critical engineering across governmental aviation infrastructure, production AI startups, and accredited professional bodies.

Software Developer Intern

Pakistan Airport Authority (PAA)

07/2026 — 08/2026 Onsite • Karachi, Pakistan

Contributed to aviation operational systems with high-reliability requirements, focusing on enterprise database integrity and end-to-end web workflows.

Developed a full-stack internal enterprise web application using ASP.NET (C#), bridging secure user interfaces with robust back-end business logic.
Architected and optimized SQL Server database schema, including relational constraints, stored procedures, and high-integrity queries for aviation records.
Collaborated within a rigorous engineering team to deliver a production-ready internal platform meeting tight deadlines.

ASP.NET (C#) SQL Server Enterprise Architecture Aviation Data Systems

ML & GenAI Engineer Intern

NeuroAGI

11/2025 — 03/2026 Remote • Applied AI

Engineered enterprise-grade Retrieval-Augmented Generation (RAG) pipelines and scalable cloud microservices for real-time document intelligence.

Designed and deployed production RAG pipelines using FAISS and ChromaDB with embedding-based semantic search for zero-hallucination enterprise document Q&A.
Engineered high-throughput RESTful microservices with FastAPI, incorporating Pydantic validation, structured telemetry, and rate limiting.
Containerized AI microservices with Docker and orchestrated deployment onto Microsoft Azure across staging and production environments.

FastAPI ChromaDB FAISS Docker Microsoft Azure RAG Pipelines

BS in Software Engineering

Iqra University

2023 — 2027 (Expected) Karachi, Pakistan

Undergraduate degree emphasizing foundational computer science, algorithm design, software architecture, and distributed intelligence.

Comprehensive study in Data Structures & Algorithms, Database Management Systems, Distributed Computing, Object-Oriented Software Design, and Artificial Intelligence.
Consistently applying theoretical principles to open-source production projects and research benchmarks.

Algorithms & Data Structures Distributed Systems Database Engineering

IBM Generative AI Engineering Professional Certificate

Coursera / IBM

Issued July 2026 Verified Professional Credential

Rigorous professional certification program validating expertise in frontier Generative AI engineering and deployment workflows.

Deep-dive curriculum covering Foundation Models, Prompt Engineering, RAG Systems, Multi-Agent Coordination, and Generative AI Application Lifecycle.

IBM Certified Generative AI Engineering LLM Architecture

06 // SECURE UPLINK

INITIATE CONTACT /

Available for Generative AI engineering roles, agentic systems consulting, and high-impact software challenges worldwide.

OPEN TO HIGH-IMPACT OPPORTUNITIES

READY TO BUILD SOMETHING EXTRAORDINARY?

Whether you are architecting a mission-critical autonomous agent pipeline, scaling zero-hallucination RAG, or building resilient AI safety layers — let's engineer the future together.

✉

Direct Email wasxy47@gmail.com

↗

☎

Phone • WhatsApp +92 313 2768054

↗

🔗

LinkedIn abdul-wasay-sheikh-xox

↗

👾

GitHub Archive github.com/wasxy47

↗