Ramis M. looks like a good fit?
We can organize an interview with Aldin or any of our 25,000 available candidates within 48 hours. How would you like to proceed?
Schedule Interview NowRamis M. Backend, Cloud and AI Platforms
My name is Ramis M. and I have over 3 years of experience in the tech industry. I specialize in the following technologies: Python, Machine Learning, Back-End Development, CI/CD, Google Cloud Platform, etc.. I hold a degree in Bachelor's degree. Some of the notable projects I’ve worked on include: Multi-AI Agents, Arthur, Jeff Booth- AI Mentor, Complain Partner, ChatGPT/GPT-4 Chatbots, etc.. I am based in Karachi, Pakistan. I've successfully completed 7 projects while developing at Softaims.
I specialize in architecting and developing scalable, distributed systems that handle high demands and complex information flows. My focus is on building fault-tolerant infrastructure using modern cloud practices and modular patterns. I excel at diagnosing and resolving intricate concurrency and scaling issues across large platforms.
Collaboration is central to my success; I enjoy working with fellow technical experts and product managers to define clear technical roadmaps. This structured approach allows the team at Softaims to consistently deliver high-availability solutions that can easily adapt to exponential growth.
I maintain a proactive approach to security and performance, treating them as integral components of the design process, not as afterthoughts. My ultimate goal is to build the foundational technology that powers client success and innovation.
Main technologies
Backend, Cloud and AI Platforms
3 years
Python
1 Year
Machine Learning
1 Year
Back-End Development
1 Year
Additional skills
- Python
- Machine Learning
- Back-End Development
- CI/CD
- Google Cloud Platform
- Amazon Web Services
- Artificial Intelligence
- Large Language Model
- Multimodal Large Language Model
- Natural Language Processing
- Microsoft Azure
- Computer Vision
- YOLO
- OpenAI API
- n8n
- Backend
- LLM
Direct hire
Potentially possible
Previous Company
Techlogix
Experience Highlights
Multi-AI Agents
Architected and deployed the World’s first grid search infrastructure multi-agent RAG systems, integrated tool-calling workflows and optimized p95 latency under 500ms across diverse LLMs using LangGraph/LangChain using AWS Bedrock KnowledgeBase.
Arthur
Developed a fine-tuned DeepSeek-R1-Distill-Llama-70B using QLoRA with Multi-GPU training on 8xH100 with FSDP, targeting reasoning in STEM and electrical engineering. Designed a QA pipeline with PyMuPDF parsing, Qwen2.5-VL OCR, DeepSeek-r1 prompts for multi-angled questions, Llama-405B judge, and retrieval via ChromaDB+text-embedding-3-large with multi-stage checkpoints. Curated 50k+ synthetic datasets (MATH, Physics, Petroleum, etc.) trained 6 epochs. Benchmarked Arthur on ElecBench & STEM-AI-mtl QA, achieving 87% accuracy, 0.61 recall, 0.46 F1, with BERT score + LLM-as-judge.
Jeff Booth- AI Mentor
Integrated AssemblyAI to transcribe and label Jeff Booth’s video content. Built a LangChain-ChromaDB pipeline to embed transcripts for contextual retrieval. Used GPT-4 Turbo with persona-tuned prompts to create a chatbot that mimics Jeff’s tone and logic. Implemented a Dropbox watcher for auto-updating the vector store. Stored session-based user memory in PostgreSQL. Deployed the chatbot as a FastAPI backend API for seamless integration. Deployed it on AWS EC2, and AWS Serverless to maintain scalability and smooth experience.
Complain Partner
Developed a FastAPI application using Python, featuring an agent built with LangGraph and powered by GPT-4o-mini as the base LLM. The application accepts PDFs, images, and text for registering complaints. Preprocessed PDFs using the Unstructured tool with OCR for better data extraction. The GPT agent utilizes custom tools and functions developed with LangChain for decision-making, interacts with third-party APIs, and determines which API endpoints to invoke based on user input. The agent serves as a complaint registrar, ensuring all prerequisites are met
ChatGPT/GPT-4 Chatbots
Objective: Develop a Python application leveraging the langchain library. This application should be capable of engaging in conversations with users, storing crucial information in a vector database, and utilizing historical and vector data to respond to inquiries. The Awesome ChatBot Project is a conversational AI assistant with both backend and frontend components. The backend uses Python, Flask, ChromaDB, Langchain, and OpenAI to enable advanced natural language processing. The frontend provides an interactive user interface built with React.JS. To set up the project, Python 3.10+ and Node.js need to be installed. The backend requires creating a virtual environment, installing dependencies like Visual Studio C++ Build Tools, and running app.py. The frontend needs installing npm packages and running npm start. Once running, the frontend seamlessly connects to the backend to deliver a robust chatbot experience. Users can have natural conversations with the AI assistant named Claude. The project utilizes cutting-edge NLP techniques to understand user input and provide relevant responses. Overall, this project enables developers to build a production-ready chatbot leveraging Python and React.JS. The README provides clear instructions to run both frontend and backend components locally. Contributors are welcome to help improve the chatbot's capabilities.
Education
FAST National University of Computer and Emerging Sciences
Bachelor's degree in Computer science
Languages
- English
Github