AI Engineer & Researcher

Abu Bakar
Siddik.

I specialize in Agentic Workflows, RAG Pipelines, and LLM Optimization. Currently engineering robust AI solutions at AskTuring.ai.

GitHub

Years Exp

+45%

Contextual Accuracy

Winner

Google Hackathon

live_eval_suite.py

CITATION ENGINE

Source Verification

SAFETY LAYER

Guardrails Active

RAG PIPELINE

Multi-Hop Reasoning

CONTEXT ENGINE

Window Optimized

WEB SEARCH

Real-time Data

VALIDATION SYSTEM

DeepEval Active

Private LLM Stack

Fine-tuning & vLLM Hosting

End-to-end pipeline: dataset curation, alignment loops, scalable inference on vLLM, and airtight observability for private deployments.

Custom fine-tunes

LoRA/QLoRA, mixed precision, guardrailed eval gates.

vLLM autoscale

Multi-replica serving with token-bucket QoS and routing.

Live Flow: Private LLM Ops

Train Serve Eval

Zoom

120%

Hover • Click • Scroll to Zoom • Drag to Pan

Agentic RAG // End-to-End

My Architecture Blueprint

Full agentic system like ChatGPT. Hover, click, and zoom into every component.

Live Flow: Agentic Reasoning Mesh

Planner Retrieval Safety

Zoom

100%

Hover • Click • Scroll to Zoom • Drag to Pan

Career Journey

The Winding Road

Navigating through complex challenges and delivering impact.

+30% Throughput

-98% Hallucinations

AI Engineer

AskTuring.ai Jul 2025 - Present

Designed multi-agent workflows improving throughput by 30%.
Built advanced agentic RAG with multi-layer memory (+45% contextual accuracy).
Reduced hallucinations by 98% via optimized grounding strategies.
Created internal eval benchmarks cutting prompt test time by 99%.

-45% Safety Risk

High Reliability

Machine Learning Engineer

Sazim Tech Ltd Oct 2023 - Jul 2025

Implemented port-and-adapter backend to support multiple LLM providers.
Improved reliability and user trust with evaluation + source-based fact checking.
Reduced AI safety risks by 45% using guardrails and advanced prompt engineering.
Deployed private AI solutions with strong data privacy and security.

Project Lead

ML Researcher & Engineer

Intelsense AI Sep 2022 - Sep 2023

Led the data annotation team as Project Lead.
Built Rasa-based chatbots for early pilots.
Collaborated with building ASR for Bengali Language.

Innovation Lab

Selected Projects

Drag to explore the archive of experiments and products.

Winner

MagicSpin 360°

Transforms one 2D image into an interactive 360° spin with Gemini multimodal.

Framework

PAGANS

Prompt normalization system for consistent prompts across LLM families.

Agent

ReleaseBot

Reads commits, drafts release notes, and emails users automatically.

Full-stack

AI Meme Generator

Natural language to ready-to-share memes with smart image selection.

Thesis

Blood Clot AI

Classifies clots and suggests treatment; thesis project.

Enterprise

SEO Keyword Optimizer

Clusters millions of keywords into themes to guide enterprise SEO.

Hall of Fame

Competitive Achievements

Top-percentile finishes across hackathons and ML contests.

Winner • Top 25 / 2,723

Google Nano Banana Hackathon

MagicSpin 360° • Generative AI • 2025

Built MagicSpin 360°: turn a single 2D subject into a seamless, rotatable view using Gemini multimodal with upload, AI-generate, or guided drawing inputs.

Winner Multimodal Gemini

Bronze Medal

Mayo Clinic STRIP AI

Kaggle • Solo • 888 teams • 2022

Stroke clot-origin classification from whole-slide pathology to guide post-stroke treatment.

Healthcare Computer Vision

9th Place

A Perfect Fit Challenge

ML Competition • 9 / 2,381 • 2021

Predictive model for selecting perfect candidate for a job description based on resume data.

NLP ML

5th Place

BioMed Datathon

AI for Medical Data • 5 / 23 • 2024

Inter-university multi-label classifier on heart-sound data for cardiac abnormality detection.

Healthcare Audio ML

6th Place

Dataverse Hack

Build AI to Save Lives • 6 / 1,145 • 2022

Computer vision system to detect potholes from roadway imagery for safer city planning.

Computer Vision Safety

Journey

Education & Milestones

Current

M.Sc. in Computer Science

Rajshahi University of Engineering & Technology (RUET)

Focus: Artificial Intelligence. Specializing in Deep Learning, Advanced Image Processing, and cutting-edge AI technologies and applications.

2019 - 2024

B.Sc. in Mechatronics Engineering

Rajshahi University of Engineering & Technology

Relevant Coursework: Artificial Intelligence, Machine Learning Algorithms, Software Engineering, Robotics, Numerical Analysis & Statistics, Digital Signal Processing & Machine Vision, Automation. Graduated with a thesis on Medical AI.

Technical Skills

Core technologies and frameworks I work with.

AI & ML

NLP & LLMs Computer Vision Agentic Workflow Fine-Tuning TensorFlow PyTorch Jax Transformers

Frameworks

LangChain LangGraph LlamaIndex Crewai smolagents Langfuse

Backend

FastAPI NestJS PostgreSQL MongoDB Weaviate Pinecone PgVector

Cloud

AWS SageMaker AWS EC2 GCP Vertex AI Docker Kubernetes GitHub Actions

Frontend

React Next.js TypeScript Python Git

Professional

Technical Writing Mentoring Agile Problem Solving

Certifications

DevOps, DataOps, MLOps

Coursera (Duke University)

Natural Language Processing Specialization

Coursera (DeepLearning.AI)

Deep Learning Specialization