- **🚀 AI/ML Engineer | Systems Engineer | Researcher |
- DXエンジニア|システムエンジニア|Web技術 (フロントエンド・バックエンド)|自動化・データサイエンスのスペシャリスト
Nice to meet you! I'm Salman Farshi, originally from Bangladesh. I’m an engineer with practical experience in デジタルトランスフォーメーション(DX), 業務プロセスの自動化, Web技術, and data science & AI.
From system development to AI implementation, I’ve been involved in a wide range of solutions to improve operational efficiency and digitalize manual processes. I strive to bring value to organizations by combining technical skills × logical thinking × integrity.
🔍 Passionate about solving real-world problems using Artificial Intelligence and Deep Learning.
💡 I specialize in Computer Vision, Generative AI, NLP, and building intelligent multi-agent systems.
🎯 Focused on contributing to cutting-edge AI research, LLM-powered applications, and open-source projects.
- North South University, Bangladesh
- Major: Computer Science and Engineering
AI/ML Engineer
- System Automation
- AI Agent
AI/ML Engineer |
- LLM-powered business automation
AI/ML Engineer
-
社内DXソリューション向けのWebシステム開発および統合に従事
-
自動化ワークフローおよびデータ分析パイプラインの構築に貢献
North South University, Dhaka, Bangladesh
- Designed and trained deep learning models for computer vision and generative AI research.
- Developed automation tools and handled data preprocessing to support research experiments.
-
LangGraph Agents Certification
Certification for hands-on experience with LangGraph Agents, focusing on building intelligent AI agents capable of complex decision-making and collaboration. -
CrewAI for Multi-Agent Systems Certification
Demonstrates proficiency in designing and deploying multi-agent systems using CrewAI, enabling scalable and cooperative AI agent networks. -
Advanced Use Cases Certification
Covers advanced applications of AI agents in real-world scenarios, emphasizing practical problem-solving with multi-agent architectures. -
LangChain for LLM Apps Certification
Validates skills in developing applications using LangChain to leverage Large Language Models (LLMs) for natural language understanding and generation. -
Functions, Tools, and Agents with LangChain Certification
Focuses on the integration of functions, external tools, and AI agents within LangChain to create dynamic and interactive LLM-powered applications.
-
Generative Models & Prompt Engineering – Skilled in designing prompts, optimizing system messages, and leveraging LLMs for creative and functional use cases
-
Task Automation with AI Tools – Built agents for routing, scheduling, email triage, and real-time task coordination
-
Custom Chatbots & Document Intelligence – Created intelligent interfaces for documents, Q&A systems, and assistant-like agents
Completed advanced AI/ML coursework through top online programs and applied learnings in real-world projects.
-
Deep Learning Specialization — DeepLearning.AI
Covered neural networks, CNNs, RNNs, and optimization techniques for building deep learning models. -
Machine Learning Specialization — Stanford University via Coursera
Studied supervised and unsupervised learning, SVMs, clustering, and applied machine learning techniques. -
Generative Adversarial Networks (GANs) Specialization — DeepLearning.AI
Hands-on experience with DCGANs, Conditional GANs, CycleGANs, and techniques to stabilize training. -
TensorFlow: Advanced Techniques Specialization — DeepLearning.AI
Focused on transfer learning, fine-tuning, TensorFlow Functional API, and advanced model training. -
MLOps Specialization (Machine Learning Engineering for Production) — DeepLearning.AI
Gained skills in deploying, scaling, monitoring, and maintaining ML models in production environments.
💡All courses were completed with hands-on practice and applied to project work.*
Here are some of my top AI & ML projects:
🛑 Note: The following projects are private due to proprietary or client-sensitive data. Demos available upon request.
Status: Completed | Tech: LangChain, LLMs, Vector Databases
- Developed a domain-specific chatbot using Retrieval-Augmented Generation (RAG) techniques.
- Integrated LangChain for memory optimization, improving response relevance by 50%.
- Focused on vector embedding, document retrieval, and conversational context handling.
- GitHub: Private | Demo: Available upon request
Status: Completed | Tech: LangChain Agents, Multi-Agent Systems, AWS Lambda, FAISS
- Built a multi-agent system for automating invoice generation, inventory management, and customer support workflows.
- Leveraged LangChain Agents to enable task delegation and execution.
- Reduced manual workload by 75%, boosting operational efficiency by 40%.
- GitHub: Private | Demo: Available upon request
- 🛰️ Deep Learning | YOLO | Computer Vision
- Detects arable land using satellite images to support agriculture.
🔹 NeuroArt - AI Art Generation
- 🎨 Deep Learning | CNN | Neural Style Transfer
- Generates artistic images using neural networks.
- 🏢 YOLO | Image Segmentation | Computer Vision
- Identifies structurally weak buildings in urban areas.
- 🌺 GANs | PyTorch | Deep Learning
- AI-generated realistic flower images using WGAN.
🔹 Deploying Custom Object Detection
- 🔍 YOLOv5 | Flask | REST API
- Deployed an object detection model with API integration.
- 🟠 TensorFlow Basics – Hands-on introduction to TensorFlow.
- 🔵 TensorFlow Advanced – Advanced deep learning models with TensorFlow.
- 🔴 PyTorch Fundamentals – Exploring the essentials of PyTorch.
- ⚫ PyTorch Advanced – Implementing deep learning architectures using PyTorch.
- 🧠 CNN Project – Implementing Convolutional Neural Networks.
- 🖌 Image Segmentation Project – Semantic segmentation using deep learning.
- 🔍 Face Detection & Verification – Face recognition and verification with TensorFlow.
- 🏗️ Basic Operations – Common data structures and algorithms in Python.
- 🎯 LeetCode Solutions – Optimized coding solutions for LeetCode problems.
| Language | Proficiency | Usage in Work |
|---|---|---|
| 日本語 | 日常会話レベル(JLPT N4) | 読み書き、メール対応、基本的な会話が可能 |
| English | Business Level | Meetings, documentation, communication |
| Arabic | Basic | All-purpose |
