JobCraft.Tech
AI job-preparation SaaS shipped solo: CV tailoring, cover letters, interview prep. 200+ active users on a freemium model.
View Details ↗
AI Consultant & Product Builder · London
I build systems that do your team's repetitive work and give you hours back every week.
Client intake, document search, internal workflows, built for small businesses. Contract-based, from a single automation to a full build.
MSc Data Science, Brunel University London · UKRI Global Talent Visa (Exceptional Talent) · Published researcher
5 yrs
Building production AI
17K+
LinkedIn followers
5
AI products shipped
UKRI
Global Talent Visa (Exceptional Talent)
Recent results
- Lead capture funnel: WhatsApp enquiries flow straight into a fully customized CRM, so every conversation is tracked and planned. More prospects now convert to clients
- Appointment reminders: one-click SMS reminders that get people through the door. 80% of appointments now show up
- Healthcare (HecoAnalytics): production RAG over research documents that cut report time by 40%
- JobCraft.Tech (my own SaaS): an AI job-preparation platform with 200+ active users
- Content engine: a fully automated video pipeline that grew a channel to 10K followers in 3 months. I ship, not just advise
Currently taking clients
Work with me
I work with small businesses where progress is stuck in emails, documents and repetitive workflows. From a single automation to a full build, no project is too small. Tell me the problem, and I'll tell you honestly whether AI is the answer.
You're in qualified hands: a data scientist with an MSc from Brunel University London, endorsed by UKRI for a Global Talent Visa (Exceptional Talent), with five years delivering production AI across legal, healthcare and finance.
Services
01 · AI Audit
Find where AI pays for itself
A fixed-scope review: your workflows mapped, savings quantified in hours and £, and a prioritised roadmap. Act on it with me or without me.
02 · Document Intelligence
Instant answers from your documents
Search contracts, reports and policies in natural language, with sources attached. The same approach cut report time by 40% for a healthcare client.
03 · Client Intake & CRM
Never miss an enquiry
WhatsApp lead capture, a CRM customized to exactly how your business works (not an off-the-shelf template), and one-click SMS reminders. In one live deployment, 80% of appointments now show up.
04 · Workflow Automation
Repetitive admin, automated
Internal tools built end-to-end, from requirements to deployment, integrated into how your team already works.
05 · Custom AI Products
From idea to production
Full product builds: LLM pipeline, backend, frontend, payments. I've shipped five of my own, including a SaaS with 200+ active users.
06 · Content & Growth Automation
Marketing that runs itself
Automated content engines and lead-generation tooling. One of mine grew a channel to 10K followers with zero manual recording.
Case Studies
80% show-up rate · Professional Services
Client intake & follow-up automation
Problem: WhatsApp enquiries were hard to track, and missed appointments were eating staff time.
Built: A CRM designed around the firm's own workflow, one-click WhatsApp capture into it, and SMS reminders.
Result: 80% of appointments now show up and every enquiry is visible in one place. Live today.
40% faster reports · Healthcare Analytics
AI reporting over research documents
Problem: Health economists assembled client reports manually from large sets of research documents.
Built: A production RAG pipeline with source-cited answers, accuracy evaluated before rollout.
Result: Report time cut by 40%; adopted across client-facing workflows.
10K followers in 3 months · Education Business
Fully automated content engine
Problem: A crowded market where SEO was unwinnable, and no time to record content manually.
Built: An automated video pipeline producing 100+ videos with zero manual recording.
Result: 10K TikTok followers in 3 months and £1,000+ in revenue, still earning passively.
How I Work
01 · Discover
A short conversation about where your team's hours actually go. If AI won't pay for itself in your case, I'll tell you that too.
02 · Roadmap
Workflows mapped, savings estimated in hours and £, priorities agreed. Fixed scope: you know exactly what you're getting before anything is built.
03 · Build
I build and integrate the system into how you already work. No rip-and-replace, and you see progress as it happens.
04 · Launch & handover
Your team trained, impact measured against the roadmap, documentation handed over. Ongoing support if you want it, with no lock-in.
Testimonials
"Very happy to recommend you after the work you did for Advanced Logic Analytics Limited in conjunction with Brunel University. The project you did for us was very well done and your skills as a Data Scientist were very good…"
Nick Ellis
Managing Director at Advanced Logic Analytics Ltd
"Abrar Is a high skilled programmer and an incisive data scientist who is capable of developing, deploying and maintaining computer models for a wide range of applications… He communicates at a high level and is able convey complex ideas simply. I am happy to recommend him."
Zaahir Dawood
Senior Data Scientist @ UserTesting
"Abrar is a skilled programmer and data scientist who works well as part of a team. Abrar acquires knowledge effectively and always demonstrates excellent work across a number of platforms (including cloud)."
David Bell
CTO at HecoAnalytics · Reader in Computer Science at Brunel University of London
AI Systems I've Built
MyUkTest.co.uk
An automated video pipeline (Polly + HTML Canvas + Flask) that hit 10K TikTok followers and £1,000+ from PDFs and courses. No manual recording, no SEO.
View Details ↗
Emotion AI for Trading
UKRI-funded research: NLP on 2M+ articles to predict stock movements. ~43% correlation achieved.
View Details ↗
Experience & Education
Feb 2026 - Present
Business Development & AI Systems @ Imran & Co Solicitors
I run the firm's content, growth initiatives, and internal software development.
- Building a custom CRM designed around the firm's specific needs
- Built a custom Lead Tracker app and Chrome extension that automatically captures WhatsApp conversations and client data, streamlining lead management
- Developed a one-click SMS reminder system, with the appointment show-up rate now at 80%
- Identify opportunities for automation and AI implementation, building internal tools that improve operational efficiency and client experience
- Own internal software development end-to-end, from requirements to deployment, for a live legal-services workflow
- Advise on growth strategy and content distribution to increase the firm's visibility and lead generation
2026 - Present
Delivery Lead @ HR Software Platform
- Leading a remote engineering team in Bangladesh building a full HR platform: clock-in/out, leave management, and related workflows
- Owning product direction, delivery and handover as the bridge between the team and the UK client
Jan 2023 - Present · Part-time
Data Scientist @ HecoAnalytics
- Built advanced RAG pipelines for health economists, reducing report generation time by 40%
- Evaluated multiple LLMs for healthcare knowledge tasks, achieving BERTScore > 0.83 and ROUGE > 0.40
- Developed AI meeting summariser converting audio to persona-specific summaries, adopted across client-facing workflows
- Automated generation of client company profiles and repetitive data ingestion tasks using LLMs and Python on AWS
- Applied topic modelling (BERTopic, LDA) and rule-based fact extraction to categorise research papers, streamlining literature review workflows
Side Activities
Jan 2026 - Apr 2026
Quantum Technology Feasibility Study · UK Style Roads
Built a voxel-based simulation modelling quantum sensor sensitivity for detecting voids, water ingress, and solid material beneath UK road surfaces.
Dec 2021 - Jan 2023
Data Scientist @ Brunel University London & Industry Partner
- Led development of a stock market behaviour analytics tool on a UKRI-funded project, using NLP on 2M+ financial news articles
- Engineered emotion analysis pipeline achieving ~43% correlation between news sentiment and stock price movement
- Took project from ideation to production in cross-functional team
May 2021 - Sept 2021
AI Intern @ CityMaas
Image-captioning models (best BLEU 0.54) to improve web accessibility for visually impaired users.
2015 - 2021
Education
MSc Data Science & Analytics, Brunel University London (2020–2021)
BSc Computer Science & Engineering, North South University (2015–2019, 87% average)
Toolbox: what I build with
AI & NLP
LLMs / GPT
RAG
LangChain
PyTorch
HuggingFace
Fine-tuning (LoRA)
Topic Modelling
OpenAI SDK
Data & Core
Python
SQL
Pandas
scikit‑learn
Cloud & Tools
AWS
SageMaker
Lambda
Bedrock
Docker
Git
CI/CD
Chrome Extensions
Research & Publications
📄
Artificial Neural Network and Machine Learning Based Methods for Population Estimation of Rohingya Refugees: Comparing Data-Driven and Satellite Image-Driven Approaches
Journal: Vietnam Journal of Computer Science
Publisher: World Scientific
DOI: 10.1142/S2196888819500246
Manual census collection is slow and costly in refugee settings, so this study explored faster alternatives: estimating the Rohingya refugee population with remote sensing and machine learning. It benchmarked 11 models, including artificial neural networks, across a data-driven and a satellite image-driven approach. Gradient Boosting performed best on the data-driven side and proved suitable for both, while the ANN performed best on satellite imagery.
View Publication🎓
Deep Learning Based Image Captioning for Visually Impaired Web Users
MSc Dissertation · Data Science & Analytics
Institution: Brunel University London (2021)
Developed an encoder-decoder image captioning system to describe website images for visually impaired users, connecting computer vision with natural language processing. Six models were built across different CNN and RNN architectures, word embeddings were compared, and the best model (BLEU 0.54) was selected for real-world implementation at CityMaas.
Read DissertationCredentials: SQL (HackerRank) · Vector Databases (Udemy) · Generative AI (Coursera) · Bloomberg Market Concepts
Common Questions
Where does our client data go?
Wherever possible I build inside your own accounts and infrastructure, so your documents never leave your control. Confidentiality constraints are scoped in writing before any build starts. I've built systems handling sensitive legal and healthcare data, so I know what's at stake.
What does an AI audit involve?
A fixed-scope review of where your team's hours actually go. I map the workflows, quantify the potential savings in hours and £, and hand you a prioritised roadmap. You can act on it with me or without me. There's no obligation beyond the audit itself.
Do we need technical staff?
No. I build, integrate, train your team and hand over documentation. If you have IT people, I'll work with them; if you don't, that's fine; most of my clients don't.
How long does a project take?
A single automation: days to a couple of weeks. A knowledge assistant: typically a few weeks. A full custom CRM: several months, delivered in phases, so you're using the first working pieces long before the whole build is done. Either way, the scope is fixed and you'll know the timeline before we start.
What kinds of businesses do you work with?
Small businesses. That's a deliberate choice, not a limitation. I've delivered in professional services, healthcare, HR and education, and small teams are where one well-built system makes the biggest difference. If your problem is repetitive knowledge work, it probably fits, whatever your industry.
Contact
Say hello
- Location: London, United Kingdom
- Email: abrar-personal@outlook.com
- LinkedIn: linkedin.com/in/abrar-fahim
- GitHub: github.com/fahimabrar
Contact Me
Working together
Whether it's a project you want built (big or small), an AI audit for your company, a collaboration, or a question about my work, the fastest way to reach me is email: abrar-personal@outlook.com. I work on a contract basis. Tell me the problem; I'll tell you honestly whether I'm the right person to ship it.