Building intelligent systems with Generative AI, Agentic AI, RAG, and LLMs. Specializing in production AI workflows, autonomous agents, and real-time AI applications.
AI Engineer passionate about building intelligent, production-ready systems
AI Engineer with expertise in Generative AI, Agentic AI, Machine Learning, and RAG systems. I build and deploy intelligent applications using modern AI frameworks including LangChain, CrewAI, AutoGen, and the OpenAI Agent SDK.
With a background in Industrial & Systems Engineering from UET Peshawar and hands-on experience across startups and remote platforms, I bridge the gap between cutting-edge AI research and real-world business solutions.
Committed to continuous learning and innovation โ from autonomous AI agents to real-time voice systems, I deliver high-impact AI solutions in fast-paced environments.
Technologies and tools I use to build AI-powered solutions
Roles where I've applied AI to solve real-world problems
AI-powered solutions that deliver real impact
Teams struggle with breaking down projects into actionable tasks, estimating time and resources accurately, and allocating work efficiently โ leading to missed deadlines and poor resource utilization.
Built a multi-agent CrewAI system with 3 specialized agents โ Project Planner, Estimation Analyst, and Resource Allocation Strategist โ that break down projects into tasks, estimate effort, and assign work to team members. Also integrated Trello board data fetching for automated sprint progress reporting using custom tools.
Diners struggle to find restaurants that match their preferences โ vibe, cuisine, or occasion โ through scattered reviews and static listings that don't understand natural language queries.
Built an intelligent conversational assistant using Groq LLM via LangChain, ChromaDB for multimodal embeddings, and a ReAct agent loop with MCP tools. Users query restaurants in natural language by name, vibe, or cuisine and receive structured recommendations powered by semantic search and agentic reasoning.
Content teams face challenges maintaining quality, ethics, and originality at scale โ manual checks are slow and plagiarism goes undetected.
Building a Generative AI Content Optimization System with integrated quality checks, ethics validation, plagiarism detection, and real-time analytics to enhance content quality while maintaining ethical standards.
Customer support teams struggle with high query volumes, inconsistent responses, and delayed ticket resolution, leading to poor customer experience.
Developed autonomous AI agents that handle customer queries end-to-end using NLP for intent detection and sentiment analysis. Enabled automated conversation management and continuous learning with AI teams.
Manual workflows cause delays, human error, and inconsistent task prioritization. Businesses need intelligent automation for ticket management and operational efficiency.
Developed multi-agent systems using CrewAI and AutoGen for autonomous task execution, planning, and operations โ reducing manual effort by 40%. Built AI workflows for ticket classification, prioritization, and escalation with 35% improvement in data-driven insights.
Existing chatbot systems lack natural conversation flow, introducing latency and losing context, making them unsuitable for real-time voice interaction and multilingual use cases.
Built an AI-powered voice-to-voice chatbot using speech recognition to convert voice to text, LLM for processing, and TTS synthesis for responses. Supports continuous dialogue, context retention, and low-latency multilingual communication.
Businesses need intelligent, context-aware chatbots that provide accurate and personalized customer interactions โ but generic solutions fail to understand domain-specific queries.
Built LLM-powered chatbots and recommendation systems with intelligent response generation. Integrated Gen AI models for contextual customer interaction. Analyzed 50K+ record datasets to identify patterns and business insights.
Where I've led teams and competed globally
Spearheaded the Polyglot Meeting Whisper Project, monitoring and guiding 20+ international participants. Ranked Top 10 globally. Delivered a multilingual AI-powered meeting transcription and summarization solution that cut post-meeting turnaround time by 25%.
Built and evaluated a PR review agent that grounds its findings in real test-suite execution rather than static diff reading. Designed a controlled eval harness across 11 PR fixtures (10 with injected bugs, 1 deliberately hard mixed refactor) and ran a head-to-head comparison against a non-tooled baseline, measuring actual improvement rather than perceived quality. The project emphasizes reproducibility, honest reporting of failure modes, and clear documentation of trade-offs โ including why a keyword-based eval heuristic was chosen over an LLM judge.
Built a full-stack heat-risk triage tool for Phoenix, AZ that ranked neighborhoods by heat urgency by combining live temperature data from the FortyGuard Temperature API with demographic vulnerability indicators (elderly population, low-income households, tree canopy). Developed a Python/Flask backend that normalized and weighted multi-source data into a tunable risk score, served it via REST endpoints, and cached API responses with graceful fallbacks. Built an interactive React + Leaflet frontend that visualized ranked zip codes on a color-coded map, including a real-time "Simulate Heatwave" animation that demonstrated how risk scores shift under extreme conditions โ the project's demo centerpiece.
Milestones and recognitions
Graduated with a B.S. in Industrial & Systems Engineering from the University of Engineering & Technology (UET) Peshawar, ranking among the top 5% of the batch.
Professional certifications that validate my expertise
I'm always open to new opportunities and collaborations
Feel free to reach out for AI consulting, collaboration on AI projects, or any exciting opportunities in Generative AI and Agentic AI.