Adil Shamim

Co-Founder

Dhaka, Bangladesh1 yr 9 mos experience

AI EnabledAI ML Practitioner

Key Highlights

Reduced AI inference time by 56% in research.
Ranked Top 1% on Kaggle out of 4,082 competitors.
Founded Toolly, a live AI tools platform.

Stackforce AI infers this person is a SaaS-focused AI Engineer with strong capabilities in machine learning and product development.

Contact

Skills

Core Skills

Machine LearningAi Tools DevelopmentData ScienceResearch

Other Skills

Chatbot DevelopmentUI/UX DesignAnalyticsData VisualizationDeep LearningMLOpsGenerative AIPython (Programming Language)Computer ScienceData StructuresDocker ProductsData AnalysisCascading Style Sheets (CSS)HTMLC (Programming Language)

About

Most ML engineers build models. Few ship systems that move real business numbers. I do both. I published a 1st-author conference paper that cut AI inference time by 56% — presented at BUET CSE Fest 2026. I ranked Top 1% on Kaggle out of 4,082 competitors. I built a recommendation system that lifted a real client's sales by 10% in 90 days — deployed, measured, verified. I founded Toolly — a live AI tools platform with 400+ tools across 15 categories, built and shipped solo at 21. My approach: production from day one. Every project I build has a deployment target, a metric, and a feedback loop. I don't consider a model real until it serves actual traffic and moves an actual number. Stack: Python · scikit-learn · XGBoost · TensorFlow · LangChain · LangGraph · ZenML · MLflow · Docker · FastAPI Currently looking for a full-time ML or AI Engineer role. Building something in AI and want a second opinion on architecture or deployment? Let's talk.

Experience

1 yr 9 mos

Total Experience

1 yr 4 mos

Average Tenure

1 yr

Current Experience

Buet cse fest 2026

Researcher — 1st Author (Conference Paper)

Feb 2026 – Mar 2026 · 1 mo

Title: Researcher — 1st Author (Conference Paper)
Company: BUET CSE Fest 2026 · DL Sprint 4.0
Date: Mar 2026
Published "Bangla Diarizz: Domain-Adapted Speaker
Diarization for Bengali Long-Form Audio"
→ Achieved 56% wall-clock speed improvement —
reduced inference time from 1h 22m to ~36m across
14 test audio files.
→ Surgically replaced segmentation and embedding
modules of Pyannote-Community-1 pipeline.
→ Fine-tuned segmentation model using diarizers
library; replaced speaker embeddings with
WeSpeaker ResNet34-LM.
→ Competitive DER on Bengali-Loop benchmark.
Read on ResearchGate: https://www.researchgate.net/publication/401194830_Bangla_Diarizz_Domain-Adapted_Speaker_Diarization_for_Bengali_Long-Form_Audio

ResearchMachine Learning

Toolly

Founder

Jun 2025 – Present · 1 yr · On-site

Most AI directories charge creators to get listed. I built one where they don't and shipped the entire product solo.
Toolly (toolly.tech) is a live, community-driven AI tools platform where creators submit their tools for free and users discover trusted AI utilities—400+ tools, 15 categories, real users, and zero paywalls.
What I built:
Architected and launched the full product from scratch—frontend, submission pipeline, moderation system, search/filter logic, and analytics—solo, in production
Engineered a free tool submission flow that onboarded early creators and built a growing contributor community with zero paid acquisition
Built and shipped Toolly Studio — a Streamlit + Bria AI image generation app with batch export and Docker deployment, bridging ML and non-technical users
Designed the full UI/UX, content strategy, and performance optimization to maximize discoverability and retention
Set up analytics and tight feedback loops to iterate rapidly—206+ commits over 10 months, the same shipping discipline as engineering teams at funded startups
Added a Learn AI educational hub directly inside the platform — because discovery without understanding doesn't move people forward
The result: A live, indexed, growing product — built by one person, used by real users, iterated based on real feedback. Not a side project. A company.

Chatbot DevelopmentMachine LearningAI Tools Development

Kaggle

Kaggle & Open-Source Contributor

Sep 2024 – Present · 1 yr 9 mos · Remote

Crafted a suite of Kaggle notebooks that transform raw data exploration, machine learning, and visualization into an intuitive, story‑driven experience.
Tackled real‑world challenges through Kaggle competitions, leveraging AI to push the boundaries of what’s possible.
Collaborated on open‑source AI and ML projects via GitHub, building tools that accelerate innovation for everyone.
Built interactive Jupyter and Colab projects designed not as static examples, but as dynamic platforms that invite experimentation, spark curiosity, and empower creators.

Data VisualizationDeep LearningData ScienceMachine Learning