Curriculum Vitae

Md Bazlur Rahman Likhon

Senior Cloud & AI Engineer

Generative AI, RAG, computer vision, conversational AI, MLOps, secure cloud-native systems, and enterprise multi-cloud architecture.

MD BAZLUR RAHMAN LIKHON

Senior Cloud & AI Engineer

Senior Cloud & AI Engineer and Head of AI Engineering with 6+ years of experience designing, deploying, and optimizing production-grade AI and cloud-native systems across AWS, Google Cloud, Azure, Oracle Cloud, and Alibaba Cloud. Specialized in Generative AI, Retrieval-Augmented Generation, LLM applications, computer vision, conversational AI, and secure multi-cloud architecture, delivering solutions for clients across Bangladesh, the USA, the UK, Japan, and China. Proven track record of leading end-to-end AI initiatives from concept to production and building scalable, secure, high-performance systems with measurable business impact. Holds a BSc in Computer Science and Engineering and is pursuing a Master’s in Information Systems Security.

Core Competencies

Generative AI & LLM Engineering (RAG, Agents, Prompt Engineering)
Computer Vision & Biometric Systems (Face Recognition, Fingerprint, YOLO, OpenCV)
AI Surveillance, Security & Behavior Analytics
Cloud Architecture: AWS, Google Cloud, Azure, Oracle Cloud, Alibaba Cloud
MLOps, CI/CD, Docker, Kubernetes, Serverless
Python, FastAPI, Flask, Node.js, SQL, C++
BigQuery, PostgreSQL + pgvector, Milvus, ChromaDB
NLP, Sentiment Analysis, Intelligent Document Processing
Conversational AI, Voice Assistants, Call Center Automation
Multi-cloud Security, Networking, Disaster Recovery

Work Experience

Senior Cloud & AI Engineer

Oct 2025 – Present

Upstra Communications Limited · Dhaka, Bangladesh

  • Led the AI engineering function with full ownership of end-to-end delivery of enterprise AI solutions spanning RAG platforms, LLM systems, conversational AI, voice assistants, computer vision, and intelligent surveillance.
  • Architected and delivered production-grade RAG systems using PostgreSQL + pgvector, Milvus, and ChromaDB, integrating frontier LLMs and open-source models with hybrid embedding strategies for cost, latency, and accuracy optimization.
  • Designed and implemented local, cloud, and hybrid LLM deployments enabling secure, low-latency, and data-residency-compliant AI platforms.
  • Developed advanced biometric systems including large-scale fingerprint matching and face recognition pipelines with high-accuracy identity verification.
  • Delivered intelligent document understanding systems for prescriptions and invoices, automating extraction, validation, and downstream processing.
  • Led AI-based surveillance, monitoring, and behavior-analysis systems including human activity recognition, anomaly detection, and intelligent alerts.
  • Partnered with major banks and pharmaceutical enterprises to design and deploy AI-driven financial and healthcare solutions at scale.
  • Led and mentored AI engineers, set architectural standards, reviewed designs and code, and ensured delivery quality, security, and timelines.
  • Acted as the technical bridge between business, clients, and engineering through solution workshops, live demos, and executive-level presentations.

Cloud Engineer

Sep 2023 – Oct 2025

Kamal-Paterson Ltd & KPCloud · Dhaka, Bangladesh

  • Spearheaded the design, development, and deployment of complex AI and SaaS applications, managed junior development teams, and oversaw the full project lifecycle from concept to production.
  • Architected the McKrishiSheba AI agricultural advisory platform using FastAPI, PostgreSQL with pgvector, Google Gemini, React + TypeScript, and Bangla TTS/STT; reduced AI costs by approximately 66–83% through token optimization.
  • Led Sleepy Owl Stories, an AI story generator delivered as a production-ready application on Google Play within four months, integrating text generation, image generation, and text-to-speech.
  • Guided Pocket Chef AI cooking assistant development, including its data-scraping pipeline and custom-trained AI model.
  • Led production and deployment of the KP Task, KP Cards, KP Space, KP Learn, KP HRM, and KP Commerce SaaS portfolio on AWS.
  • Managed critical security incidents, including hacked WordPress recovery, a 14-day enterprise email disaster recovery effort, and multiple workload migrations to AWS.
  • Contributed to cloud projects in the USA, UK, Japan, and China, adapting to diverse technical environments.

Cloud Trainer

Aug 2023 – Nov 2023

Inflexionpoint Technologies BD Ltd. · Dhaka, Bangladesh

  • Led a team of 20 interns and provided hands-on training on AWS, Microsoft Azure, cloud architecture, web development, machine learning deployment, and API testing.
  • Increased trainees’ cloud-platform knowledge by 80% through pre- and post-training assessments and achieved a 90% transition rate to cloud-related projects.
  • Designed and delivered a hands-on lab environment for machine learning deployment on the cloud, improving trainee project-completion rates by 25%.

Researcher

Jun 2020 – Aug 2023

DIU NLP & ML Research Lab, Daffodil International University · Dhaka, Bangladesh

  • Conducted advanced research in Natural Language Processing and Machine Learning, contributing to innovative algorithms and models.
  • Managed and contributed to a project focused on making the Bengali language more computation-friendly.
  • Collaborated with researchers, computational linguists, and data scientists on multidisciplinary AI projects.
  • Scraped and analyzed 2.4 million tweets for sentiment analysis on the Ukraine–Russia war and co-authored a published classification framework.

Key Projects

GuardZone – AI Guard Monitoring System

AI-powered security guard monitoring across multiple locations, including attendance, presence verification, movement tracking, prolonged-absence detection, real-time alerts, and audit logs.

Conversational Analytics Platform

Enterprise platform that lets business users explore organizational data through natural language, generating insights, SQL queries, interactive visualizations, and contextual follow-up recommendations in a secure session-based interface.

WorkStation Monitoring – AI Employee Behavior Analytics Platform

Real-time workplace monitoring for occupancy, activity, work status, posture, productivity, live dashboards, behavioral insights, performance KPIs, and event-based audit trails.

LunchGate – AI Food Distribution Tracking System

Facial-recognition-based cafeteria access and meal-distribution platform with shift-policy enforcement, duplicate prevention, transaction auditing, and instant violation alerts.

FaceTrack AI – Intelligent Attendance Management System

Production-grade face-recognition attendance platform with age-invariant matching, anti-spoofing liveness detection, multi-camera IP support, a role-based admin dashboard, 28+ REST APIs, enterprise controls, GPU acceleration, and cloud readiness.

KYC Face Match Platform

Banking-grade, multi-tenant KYC face verification platform with tenant isolation, role-based administration, usage tracking, billing controls, asynchronous processing, short-lived tokens, certificate-based access, audit trails, revocation flows, and tamper-evident logging.

Call Center AI – Bangla-First Voice Automation

Outbound AI call-center prototype using FastAPI, Twilio Media Streams, and Gemini 2.5 Flash Native Audio, with a low-latency WebSocket bridge, Bangla-first conversations, inquiry routing, and enterprise-grade controls.

RAG Online Shop – AI Retail Assistant (Vertex AI)

Production-ready PC-hardware recommendation system using Node.js, PostgreSQL + pgvector, Docker, Vertex AI, 768-dimensional embeddings, IVFFlat indexing, Gemini 2.0 Flash, auto-embedding triggers, sub-second responses, and approximately 80%+ relevance accuracy.

Document Processing Engine

Python Flask platform using Google Cloud Document AI and OpenCV for blur detection, multi-image PDF conversion, structured schema extraction, line-item table generation, and drag-and-drop processing.

Network Intrusion Detection (CIC-IDS-2017)

Cybersecurity data-science pipeline on 2.8M+ records using 12 machine-learning algorithms; XGBoost achieved 99.87% accuracy and 99.99% ROC-AUC, with SHAP/LIME explainability, forensic timelines, and PDF reporting.

Professional Fingerprint Processing System

Biometric web application with fingerprint enhancement, ORB + Gabor filtering, quality scoring, WebRTC capture, enrollment and verification panels, SSL security, and AWS EC2 deployment.

Dense Crowd Head Counter – GCP Deployment

Video-based unique-person counting solution for corridor-style spaces with resilient background jobs, annotated-video downloads, detailed CSV logs, and one-command GCP deployment.

NID OCR System

REST API extracting 10 structured fields from Bangladesh National ID cards using regex parsing and neural machine translation, supporting multipart and base64 uploads with timestamped audit JSON responses.

Mushok Processing Engine

Automated extraction of 23 fields from Bangladesh Mushok 6.3/VAT invoices, including buyer and seller information, line items, and VAT calculations, with magic-byte validation, blur pre-screening, and multi-image PDF merging.

Medical Prescription Extraction & Medicine Matching

Structured clinical-data extraction from handwritten and printed prescriptions using schema-constrained generation, multi-hypothesis medicine matching, batch processing, web and CLI interfaces, and JSON/CSV/ZIP export.

Face Similarity AI – Cloud

Production-grade face-comparison service with cloud AI APIs, secure server-side credentials, Node.js, Nginx, a drag-and-drop interface, and automated Debian Linux setup.

Tencent Hunyuan 3D Global Python Client

Python client and web interface for Tencent Hunyuan 3D Text-to-3D and Image-to-3D generation, including asynchronous jobs, automatic polling, configurable modes, credit estimation, and PBR materials.

InsureCow – Cattle Muzzle Detection

Machine-learning model for cattle muzzle detection and identification, plus migration to an AWS Lambda serverless architecture for improved performance and cost efficiency.

Sentiment Analysis on Ukraine–Russia War

Scraped and analyzed 2.4 million tweets and developed a classification framework for identifying public support perspectives in the conflict.

Bike Sharing Demand Prediction

AWS ML Scholarship project predicting hourly bike-rental demand using AutoGluon, feature engineering, hyperparameter tuning, and Kaggle integration.

Awards & Recognition

Judge – THE INFINITY AI BUILDFEST 2026

Evaluated 208 AI solutions for production readiness, technical excellence, and real-world impact.

VCP Champion – THE INFINITY AI BUILDFEST 2026

Recognized for building a production-ready AI solution within one hour.

Top 50 Achiever – Google Cloud Gen AI Academy APAC Edition

Apr 2025 — Recognized among the Top 50 achievers for building a production-grade Generative AI project on Google Cloud.

AWS AI & ML Scholar '24 – Udacity / AWS

Jun 2024 — Awarded the AWS AI & ML Scholarship, valued at $4,000, for the AI Programming with Python Nanodegree.

Top Contributor – Google Cloud Community 2022

Recognized for actively sharing solutions and contributing to the global community.

Next Generation Tech Booster Scholarship – Bertelsmann

Dec 2023 — Awarded for promise in AI, machine learning, data science, and cloud computing.

AWS Machine Learning Scholarship

Oct 2021 — Completed an intensive program covering SageMaker, Glue, Lambda, and model deployment at scale.

Education

Master of Science in Information Systems Security (MISS)

Dec 2024 – Present

Bangladesh University of Professionals (BUP) · Dhaka, Bangladesh

Bachelor of Science in Computer Science and Engineering

Sep 2019 – Sep 2023

Daffodil International University · Dhaka, Bangladesh

GPA: 3.62/4.00 · Thesis: Sentiment Analysis on Ukraine–Russia War Using Tweets

Selected Certifications

Google Cloud Certified Professional Cloud DevOps Engineer
Google Cloud Certified Professional Machine Learning Engineer
Google Cloud Certified Professional Data Engineer
Google Cloud Certified Professional Cloud Database Engineer
Google Cloud Certified Professional Cloud Security Engineer
Google Cloud Certified Professional Security Operations Engineer
Google Cloud Certified Associate Cloud Engineer
Google Cloud Certified Generative AI Leader
Google Cloud Digital Leader
Microsoft Certified: Azure Developer Associate (AZ-204)
Microsoft Certified: Azure AI Engineer Associate (AI-102)
Microsoft Certified: Fabric Data Engineer Associate (DP-700)
Oracle Cloud Infrastructure 2024 Generative AI Certified Professional
Oracle Cloud Infrastructure 2025 Certified AI Foundations Associate
AI Programming with Python Nanodegree — Udacity / AWS
Professional Certificate in Kubernetes and Cloud Native Technologies — The Linux Foundation
Professional Certificate in DevOps: Practices and Tools — Columbia University / edX
Intel Edge AI Certification
Proofpoint Certified AI Data Security Specialist 2025
GitHub Foundations

Technical Skills

AI & Machine Learning: Generative AI, RAG, LLM, Prompt Engineering, Computer Vision, NLP, Sentiment Analysis, YOLO, OpenCV, SFace, PyTorch, TensorFlow, Keras, Scikit-learn, AutoGluon, Deep Learning, CNN, LSTM, Reinforcement Learning

Cloud Platforms: Google Cloud Platform, AWS, Microsoft Azure, Oracle Cloud Infrastructure, Alibaba Cloud, Tencent Cloud

Databases & Vector Stores: PostgreSQL, pgvector, MySQL, BigQuery, Milvus, ChromaDB, SQL

DevOps & Infrastructure: Docker, Kubernetes, CI/CD, CircleCI, Jenkins, GitLab CI, Travis CI, Linux, Nginx, Apache, AWS EC2, Cloud Run, Firestore, AWS Lambda, Git, GitHub

Frameworks & Languages: Python, FastAPI, Flask, Node.js, C++, REST APIs, WebSocket, WebRTC

Data & Analytics: BigQuery, Power BI, Pandas, NumPy, SHAP, LIME, Data Warehousing, Data Lakes

Security: Enterprise-grade authentication, SSL/TLS, token-based security, audit trails, compliance systems, AI data security

Automation & Low-code: n8n workflow automation, Loveable AI low-code, CalDAV integration, API orchestration

Languages

Bengali — Native English — C2 Proficient German — A2 Basic

Volunteering & Leadership

Student Ambassador – Hult Prize · Sep 2019 – Dec 2019

Daffodil International University
Organized the university-level Hult Prize event with the highest number of participating teams in DIU history, mentored team formation, and managed information dissemination.

Lab Prefect – OOP II (CSE233) · Jan 2021 – Apr 2021

Daffodil International University
Managed Python and OOP laboratory sessions, guided practical implementations, and resolved technical challenges.

Lab Prefect – Data Mining & ML Lab (CSE322) · May 2021 – Dec 2021

Daffodil International University
Oversaw the data-mining and machine-learning lab, supported students applying complex concepts, and collaborated with faculty.

References

References available upon request.