AI/ML Engineer - Real-time RAG Systems & LLM Solutions expert

EDUCATION AND TRAINING

Bachelors in Computer and Information Systems Engineering (Washington Accord Accredited - PEC Pakistan Signatory)

NED University of Engineering and Technology, Karachi, Pakistan [ 27 Oct 2020 – 29 Oct 2024 ] City: Karachi Country: Pakistan Website: https://www.neduet.edu.pk/ Field(s) of study: Computer Systems Engineering Final grade: 3.744 / 4.000

Cambridge International A-Levels

Nixor College, Karachi, Pakistan [ 1 Jun 2018 – 1 Jun 2020 ] City: Karachi Country: Pakistan Website: https://www.nixorcollege.org/ Field(s) of study: Pre Engineering Final grade: Physics / Chemistry / Maths (A / A*/ A) Cambridge International O-Levels / IGCSE

New World Int’l School, Al-Khobar, KSA / The City School Satellite Town, Rawalpindi, Pakistan [ 1 Jun 2015

– 1 Jun 2018 ]

City: Al-Khobar Country: Saudi Arabia Website: https://www.newworldschool.com.sa/ Field(s) of study: Science Final grade: Physics / Chemistry / Maths / IT (A*/A*/A*/A*), English / Urdu / Islamiat / Pak Studies (A/A/A/ A)

WORK EXPERIENCE

Jaffer Business Systems (Private) Limited – Karachi, Pakistan AI/ML Engineer I

[ 22 Nov 2024 – Current ]

● Developed and deployed a real-time SharePoint RAG system on AWS EC2 with Streamlit UI, FastAPI webhooks, ChromaDB indexing, and secure Nginx HTTPS.

● Built and deployed a Salesforce MCP conversational assistant on Railway with OpenRouter LLM integration and Salesforce API tools for natural language CRM data querying and navigation.

● Built and enhanced an AI assisted insurance document processing workflow (React/TypeScript, FastAPI, Google Gemini LLMs, EZLynx, AWS S3) with robust upload, parsing, matching, and filing capabilities.

● Designed and fine-tuned LLMs (LLaMA-3, Mistral, GPT4-OSS, GPT-4o-mini) for decoding electronic component markings into structured metadata, boosting accuracy with data cleaning and JSONL training.

● Developed an MPN code semantic RAG search system with hybrid exact-match + semantic retrieval using ChromaDB, Hugging Face embeddings, and OCR normalization.

● Built a secure multi-LLM chatbot (ChatGPT + Gemini) with Google OAuth login, chat history management, and Google Drive export functionality.

● Created AI Agents with Synthflow.ai and n8n automation tools, including fine-tuning OpenAI models and automating calling workflows based on user preferences. Daniyal Amir

Passport: BE1228923 ID: 422********** Work permit: Pakistani Nationality: Pakistani Date of birth: 27 Jan 2003 Place of birth: Karachi, Pakistan

Phone number: +92-336*******

Email address: ***************@*****.***

LinkedIn: Daniyal Amir

Github: Daniyal-27

Home: House No B-121, Gulshan-e-Iqbal, Block 10, 75300 Karachi (Pakistan)

● Developed and deployed an automated n8n workflow integrating Google Sheets with Gmail and Drive, streamlining project data processing, email alerts, and centralized storage.

● Fine-tuned Stable Diffusion (LoRA) on a custom slippers dataset with successful inference.

● Fine-tuned YOLOv5 for rooftop detection from aerial imagery, achieving improved accuracy.

● Utilized Google Maps API to measure roof areas and analyze structural components.

● Developed an interactive AI application using Bubble.io (frontend + backend), integrating plugins and Gemini/ OpenAI models for personalized data generation.

● Designed an Open Perfect Hash Function to meet custom requirements.

Shispare (Private) Limited – Karachi, Pakistan

AI/ML Apprentice

[ 12 Aug 2024 – 15 Nov 2024 ]

● Employed deep learning and computer vision techniques for action recognition and video classification tasks, improving accuracy by 15%.

● Implemented SpeechToText Translator leveraging WhisperX pre-trained model, efficiently handling alphanumeric cases.

● Utilized deep learning techniques for prompt classification tasks.

● Extracted meaningful insights from database and created interactive web based visualizations using Plotly library and Dash framework.

Unikrew Solutions (Private) Limited – Karachi, Pakistan ML Internee

[ 4 Sep 2023 – 3 Oct 2023 ]

● Collaborated with the Mobile Development Team to design and implement an Android Card Scanner using the OpenCV library.

TPS WorldWide – Karachi, Pakistan

DevOps Internee

[ 1 Mar 2023 – 2 May 2023 ]

● Learned Selenium Automation with C# including the Nunit Method.

● Successfully automated tasks including Google Search functionality and tests on the Selenium test platform. SKILLS

Python / Torch, Keras, Tensorflow, Matplotlib, Numpy, Scikit-learn / Hugging Face transformers / Jupyter (Jupyter- lab/Jupyter-notebook) / Google Collaborator / Frameworks and Libraries: FastAPI, Open CV, Tensorflow, Keras, Pandas, Seaborn, Plotly, Streamlit) / GCP, AWS-S3, AWS-EC2, AWS-Lambda / Ngrok / n8n / MySQL / Bubble.io / Nginx / MATLAB / StarUML / MS Office / Docker / MongoDB / Github / Postman / PostgreSQL / VS Code / Make.com / Railway / Hugging Face Hub / DBeaver

PROJECTS

SafeNav – Final Year University Project

• This project aimed to visualise crime data on Google Maps, enabling the identification of crime hotspots in Karachi for enhanced situational awareness and decision-making, using both a mobile application and an admin dashboard panel.

• Leveraged Machine Learning Models to predict crime score for each location with respect to time.

• Technologies Used: Python, Flutter, Dart, React, MaterialUI, Flask, Render, Express, Scikit-Learn, Jupyter Notebook

Other language(s):

English ( IELTS ACADEMIC - Overall Band Score 8.0 ) LISTENING C1 READING C1 WRITING C1

SPOKEN PRODUCTION C1 SPOKEN INTERACTION C1

CIFAR-10 Image Classifier - Lab Project

• Trained a Convolutional Neural Network (CNN) to classify CIFAR-10 images utilizing the Keras Sequential API.

• Technologies Used: Python, TensorFlow, Keras, Google Collab MultiClass Weather Classifier - Lab Project

• Designed and implemented a machine learning system to classify outcomes based on weather data for predictive analysis.

• Technologies Used: Python, Scikit-Learn, CatBoost, Jupyter Notebook Water Jug App - Lab Project

• Developed an Artificial Intelligence (AI) based Water Jug Problem application for mobile phones using Android Studio.

• Technologies Used: Android Studio, Java

LANGUAGE SKILLS

Mother tongue(s): Urdu

Levels: A1 and A2: Basic user; B1 and B2: Independent user; C1 and C2: Proficient user

Contact this candidate