I design and ship production-grade AI applications — autonomous agents, RAG pipelines, and LLM-powered products that solve real problems in healthcare, education, and research. I turn cutting-edge models into reliable, deployable systems using LangChain, LangGraph, Python, and FastAPI.
From research papers to production agents — how I turn modern AI into products people actually use
I'm Ahmed Iqbal, an AI Engineer focused on Large Language Models, Agentic AI, and Retrieval-Augmented Generation. I recently completed my B.S. in Software Engineering (CGPA 3.93/4.0) and now build end-to-end AI systems that move beyond demos into real, deployable products.
What I do: I architect and ship LLM-powered applications — autonomous agents with human-in-the-loop control, RAG pipelines grounded in private data to eliminate hallucinations, and clinical and educational tools used in real workflows. My stack centers on Python, LangChain, LangGraph, FastAPI, vector databases, and modern model APIs (OpenAI, Groq, Gemini).
Proven impact: As a freelance AI Engineer I've delivered production systems for paying clients across healthcare, sports, and e-commerce. During my data science internship at Cognifyz Technologies I improved model accuracy from 87% to 92% through better feature engineering and evaluation, and a RAG system I built cut lecture-revision time for users by roughly 50%.
What I'm after: AI-first teams building agentic and LLM-driven products. I care about correctness, grounding, and getting models into users' hands — not just notebooks.
The tools and techniques I use to design, build, and ship production AI systems
Proficiency levels in key technologies and tools
Breaking down complex problems into manageable solutions
Effective communication and teamwork in agile environments
Creative approaches to data science challenges
Quick adaptation to new technologies and methodologies
Production AI systems — agents, RAG pipelines, and LLM-powered products built for real users and paying clients
Autonomous LangGraph agent that monitors Google Classroom in real time, generates solutions to new assignments with LLMs, and submits them through Google Drive. Built a human-in-the-loop layer so users approve, reject, or regenerate any response before submission — combining full automation with reliable oversight.
AI-powered clinical decision support tool (web & desktop) that analyzes patient records, lab results, and prescribed medications to generate drug-interaction, side-effect, and risk-assessment reports for pharmacists. Shipped with a centralized admin platform for pharmacy management, user monitoring, analytics, and subscription control.
RAG system that lets users ask questions about any YouTube lecture and get answers grounded strictly in the video's content — eliminating hallucinations. Built the full pipeline: video transcription, semantic search over a vector store, and LLM-based answering. Cut lecture-revision time for users by roughly 50%.
Web platform that verifies the authenticity of research papers by detecting AI-generated and human-written text, and identifying manipulated images. Combines ML models for text analysis with CNNs for image classification, plus LLM-based summarization that produces structured integrity reports.
Conversational health assistant that answers user questions from trusted medical documentation. Orchestrated with LangChain and a RAG pipeline over GPT models so responses stay grounded in source material rather than the model's open-ended generation.
AI coaching assistant for physical trainers that generates personalized training and recovery schedules, helping athletes balance practice and rest for optimal game-day readiness. Parses uploaded GPX run data, returns performance feedback, and serves recommendations through secure dashboards.
Web-based media tool that automatically corrects shaky footage and upscales low-quality images for client media workflows. Built the computer-vision processing pipeline with OpenCV and exposed it through a clean web interface for non-technical users.
Automated web-scraping platform that aggregates live sales from leading apparel, footwear, and home-goods brands into a single storefront, then redirects shoppers to the source sites to complete purchases. Engineered resilient scrapers and a unified product feed.
Content-based recommendation engine using cosine similarity over engineered movie features to surface relevant titles in real time. Deployed as an interactive Streamlit web app backed by a full movie database.
Binary text-classification model that flags spam with high accuracy on the benchmark test set. Built a complete NLP pipeline — tokenization, stemming, and TF-IDF vectorization — and shipped it as a live Streamlit app.
LLM-powered study assistant tuned for education: students work through doubts and confusing concepts in a guided, tutor-style conversation rather than receiving generic chatbot answers. Orchestrated with LangChain and LangGraph for structured, multi-step responses.
Highlighting notable accomplishments and milestones in my academic and professional journey
Graduated in Software Engineering with a 3.93 CGPA, with coursework spanning AI, NLP, DBMS, algorithms, and statistics.
Improved ML model accuracy from 87% to 92% during my Cognifyz internship through better feature engineering, preprocessing, and evaluation.
Built a RAG system over YouTube lectures that gave grounded, hallucination-free answers and cut users' revision time by roughly half.
Earned certifications from IBM (Data Science, Data Analyst), Google (Advanced Data Analytics), and Stanford / DeepLearning.AI (Machine Learning).
Delivered live AI products to paying clients — an athlete coaching assistant, a computer-vision media enhancer, and a multi-brand shopping aggregator.
Built LLM, agentic, and RAG systems spanning healthcare, education, and research — from autonomous LangGraph agents to clinical decision support.
Recognized credentials that validate my expertise in data science and AI technologies
Comprehensive certification covering Python fundamentals, data structures, and libraries essential for data science and AI development.
Advanced certification in data manipulation, analysis, and visualization using Python libraries like Pandas, NumPy, and Matplotlib.
Professional certification in data analysis, visualization, and dashboarding — covering Python, SQL, and the analytics workflow from raw data to insight.
Specialized certification in database management, SQL querying, and data extraction techniques for data science applications.
Foundational course in supervised machine learning algorithms, model evaluation, and fine-tuning for higher accuracy.
My professional journey in data science, AI engineering, and continuous learning
Design and ship production AI systems for clients across sports, media, and e-commerce — from LLM-powered assistants to computer-vision tools.
Worked end-to-end on real-world datasets, from data cleaning through model deployment in a web interface.
Completed BCG X's data science simulation, solving business problems with structured datasets and analytical reasoning.
Graduated with a CGPA of 3.93/4.0. Core coursework in AI, NLP, operating systems, DBMS, networks, algorithms, calculus, and statistics & probability.
I help teams turn modern AI — LLMs, agents, and RAG — into reliable products that ship. Whether it's a full-time role or a focused build, let's talk about what you're working on.
Open to AI/ML engineer roles and freelance builds. Reach out and let's talk.
I'm always open to discussing AI/ML engineering roles, agentic and RAG projects, or interesting problems to build with LLMs. Feel free to reach out.