Awais Mazahir

The developed system uses a single-shot multi-box detection, an SSD algorithm for localization, and a customized lightweight deep convolutional neural network as a backbone network to classify defects generally found in industrial pipes and gun barrels. First, the image dataset is acquired from a real-time working environment using an indigenously developed borescope featuring a rotating camera and special hardware features. Later, the bounding boxes are calculated using extracted features to localize defects with SSD, which takes a single shot to detect multiple objects within the image. Finally, the defected regions are classified into five classes of commonly found issues of pitting, chipping, rusting, dirt, and thermal cracking by utilizing deep learning architecture of 53 layers. It is found that the proposed approach can indicate the exact location of the classified defect in terms of angle and distance from a reference point.

A Django-based web application used to search business websites on the web-based on keywords and other AI-based search filters. It crawls the web to find the contact page, fetches the form fields, and submits the form with an automated data-filling process. The tool can send multiple contact emails simultaneously.

A Python desktop application was developed to train a custom neural network with a single click. Said algorithm was able to do binary as well as multiclass classification. The project also involves the development of a fluent UI on PyQt5, which shows live training progress.

A complete IoT Solution was developed over AWS. The system consists of a proof of concept for public eCommerce delivery cabinets which can be unlocked with face detection. It had a Python-based front end that used the Boto3 library to store data in AWS DynamoDB through AWS IoT MQTT Broker after verifying the face through Amazon Rekognition. The system used Raspberry Pi for high-end processing and Arduino for necessary actuations.

A demonstration of the prototype can be seen in the following URL:
https://www.dropbox.com/s/pm2wyvf3ip9m1ub/Smart%20Locker.mp4?dl=0

A debugging project for a US-based client's Flask web application, NurseAccred AI. The project involved identifying and resolving functional and UI/UX issues, such as subscription selection, missing links, and email validation. The fixes enhanced the website's usability and ensured smooth operation for end users.

A web application developed to leverage Amazon Affiliate API and eBay API and display the available products on search engines with Affiliate links connected to their buttons. Furthermore, the descriptions of products and their reviews were summarized using AI.

Legal AI Companion is an advanced AI-powered tool designed to assist legal professionals with document analysis, contract review, and legal research. It leverages natural language processing (NLP) to extract key clauses, identify risks, and provide case law references, streamlining legal workflows. An intuitive interface allows lawyers to quickly analyze agreements, draft legal documents, and ensure compliance, enhancing efficiency and accuracy in legal practice.

Developed the TRIPASIA AI Travel Booking Assistant, an intelligent conversational booking system integrating Google Dialogflow, OpenAI, and Node.js with Python for back-end logic. The assistant enables seamless hotel, flight, and activity searches, leveraging real-time travel APIs (Trivago, Aviasales, Expedia, Amadeus). It processes natural language queries, maintains context-aware conversations, and dynamically interacts with multiple APIs to provide accurate, real-time booking options. The system efficiently handles multi-step queries, refines user inputs, and presents results through an intuitive front-end interface. Optimized for speed, reliability, and scalability, this AI-powered assistant enhances the travel booking experience.

Developed an AI-based Vaccination Advisor using the MERN stack with fine-tuned OpenAI models, trained on 20,000+ vaccination-related responses. The system provides accurate, context-aware vaccination advice, ensuring medical and regulatory standards compliance. It features a user-friendly web interface, enabling seamless interactions, real-time query resolution, and personalized vaccination recommendations. The application provides data privacy, secure API communication, and scalability, making it a robust AI-driven healthcare advisory tool.

Developed a comprehensive ERP system for White Castle, a fast-food chain, using the MERN stack (MongoDB, Express.js, React, Node.js). The system provides multi-branch management, finance tracking, POS integration, CRM, delivery tracking, kitchen process monitoring, and inventory management. Designed for scalability and efficiency, it streamlines operations, enhances customer engagement, and optimizes financial workflows, ensuring seamless management of fast-food business operations.

I architected and developed Genesis AI, a comprehensive agentic AI platform that seamlessly integrates with smart wearables, including AR glasses and smartwatches. The platform captures real-time health metrics, environmental surroundings, and conversational data from users throughout their day, storing it in a structured data pipeline.

I built a multi-agent orchestration engine using LangGraph and LangChain to process ingested data and generate personalized "quests" — actionable daily goals tailored to the individual's health patterns, behavioral trends, and personal objectives. The system leverages RAG architecture with vector databases to maintain long-term user context and provide increasingly personalized recommendations over time.

The back end was built with Python and FastAPI, serving a React/TypeScript front end with real-time WebSocket connections for live wearable data streaming. I deployed it on AWS using containerized microservices with Docker, enabling horizontal scaling to support concurrent users. I integrated the OpenAI and Claude APIs for natural language understanding and quest generation, with custom fine-tuned models for health data interpretation.

I built SalesAICE, a fully autonomous AI sales agent SaaS platform designed for remodeling and home service professionals. The platform handles complete sales conversations from prospecting to closing without human handoff, operating 24/7 with sub-1-second response latency.

I engineered the core AI engine with advanced objection handling using tactical empathy frameworks, real-time emotion detection from vocal cues, and adaptive conversation strategies. The system integrates with existing CRM platforms for automatic lead scoring, follow-up scheduling, and deal pipeline management.

I developed a multi-tier SaaS architecture (Starter, Core, Pro, Enterprise) with SDR automation, inbound call and message AI, closer AI, nurturing agent AI, and team trainer AI modules. It was built with Next.js and React for the front end, Python and FastAPI for back-end AI services, and OpenAI for natural language processing. The platform supports unlimited concurrent conversations with consistent quality, multi-channel integration (SMS, chat, voice), and enterprise-grade compliance and data governance.

I developed MySpeakScore, a personalized AI-powered public speaking confidence assessment platform. Users submit a 60- to 90-second video and receive a comprehensive SpeakScore (0-100), a detailed analysis of their speaking strengths and focus areas, and a tailored 14-day improvement plan.

I built the AI scoring engine that analyzes multiple speaking dimensions, including clarity, confidence, structure, presence, and delivery patterns from video submissions. The system uses natural language processing and audio analysis to evaluate vocal cues, filler words, pacing, and overall communication effectiveness.

I developed the full-stack application with React for the front-end interface, Python and FastAPI for back-end video processing and AI analysis services, and integrated OpenAI models to generate warm, supportive feedback with human oversight for quality assurance. The platform supports multiple use cases, including everyday confidence-building, career preparation (interviews, presentations), and refinement of professional speaking skills. It was deployed on AWS, with secure video handling and processing pipelines.

I developed Zlayt AI for Zinda Law Group, an AI-powered legal document generation system that automates the creation of complex legal documents. The platform leverages large language models and RAG architecture to generate accurate, context-aware legal documents based on case data, client information, and jurisdiction-specific templates.

I built an intelligent document pipeline that processes case intake data and automatically drafts legal documents, including demand letters, complaints, and settlement agreements. The system uses vector databases to store and retrieve relevant legal precedents, statutes, and firm-specific templates, ensuring generated documents maintain legal accuracy and consistency with the firm's standards.

I developed the full-stack application using Python and FastAPI for the AI back end with LangChain for LLM orchestration, React and TypeScript for the front-end interface, and PostgreSQL for structured data storage. I integrated OpenAI GPT models for natural language generation with custom prompt engineering to maintain legal terminology and document formatting. It was deployed on AWS with enterprise-grade security to handle sensitive legal data and meet HIPAA-adjacent compliance requirements.