You will get Recommendation System | AI Engineer | Python developer | AI Automation
Top Rated
Project details
Unlock the potential of personalized experiences with a Machine Learning-driven Recommendation System tailored to your business needs. With over 6 years of experienced AI/ML developer, I specialize in designing, training, and deploying recommendation engines that enhance user engagement, optimize operations, and drive measurable results.
Whether you're looking for LLM-based solutions, neural networks, or advanced algorithms leveraging reinforcement learning, I bring cutting-edge expertise to every stage of the project. From data analysis and model training to fine-tuning and seamless integration, my solutions are crafted to align perfectly with your goals.
I’ve successfully built recommendation systems for industries like fintech, retail, and iGaming, optimizing customer journeys, improving conversions, and increasing ROI. My approach combines deep understanding of user behavior with state-of-the-art tools like generative AI and transformer models to create scalable, high-performing systems.
Let’s collaborate to build a solution that transforms your data into actionable insights, delivering personalized recommendations to delight your customers and grow your business! 🚀
Whether you're looking for LLM-based solutions, neural networks, or advanced algorithms leveraging reinforcement learning, I bring cutting-edge expertise to every stage of the project. From data analysis and model training to fine-tuning and seamless integration, my solutions are crafted to align perfectly with your goals.
I’ve successfully built recommendation systems for industries like fintech, retail, and iGaming, optimizing customer journeys, improving conversions, and increasing ROI. My approach combines deep understanding of user behavior with state-of-the-art tools like generative AI and transformer models to create scalable, high-performing systems.
Let’s collaborate to build a solution that transforms your data into actionable insights, delivering personalized recommendations to delight your customers and grow your business! 🚀
Machine Learning Tools
Amazon SageMaker, ChatGPT, GPT-3, NLTK, OpenCV, pandas, Python, Python Scikit-Learn, PyTorch, scikit-learn, SciPy, SQL, TensorFlow, Word2vec, XGBoostWhat's included
| Service Tiers |
Starter
$200
|
Standard
$8,000
|
Advanced
$12,000
|
|---|---|---|---|
| Delivery Time | 1 day | 40 days | 70 days |
Number of Revisions | 0 | 2 | 2 |
Number of Model Variations | 0 | 2 | 2 |
Number of Scenarios | 0 | 1 | 1 |
Model Validation/Testing | - | ||
Model Documentation | - | ||
Data Source Connectivity | - | ||
Source Code | - |
13 reviews
(13)
(0)
(0)
(0)
(0)
This project doesn't have any reviews.
VS
V S.
Jun 30, 2026
AI Automation Specialist / AI Engineer – AI Automation n8n | AI Agent Automation | Machine Learning
JW
Joern W.
Apr 3, 2026
AI Automation Specialist / AI Engineer – AI Automation n8n | AI Agent Automation | Machine Learning
Bohdan delivered a fully functional and well-documented AI pipeline with Docker-based training and inference. Communication was clear, execution was structured, and he went the extra mile to ensure a robust, production-ready result. We would gladly work with him again
EH
Emily H.
Feb 12, 2026
AI Automation Specialist Needed for Gumloop/N8N/Make Projects
Bohdan is a true AI expert - and went the extra mile to provide video walkthroughs and edit the automation to fit our needs. I would certainly recommend Bohdan to any business looking to build AI agents to help them save time and scale their business in the age of AI.
JW
Joern W.
Jan 30, 2026
Training + API + Cadify UDF Integration of dataset of 17 830 sprocket drawings
Bohdan completed on time with a super result, a fully functional Docker package.
Additionally, he was always avalable to help us with follow-up questions after that the project was actually completed and closed.
Additionally, he was always avalable to help us with follow-up questions after that the project was actually completed and closed.
RM
Raed M.
Jan 21, 2026
n8n Automation Expert: Build AI Ad Generation System
About Bohdan
AI & CRM Automation Engineer | n8n, Clay | GTM & Outreach Automation
100%
Job Success
Kyiv, Ukraine - 8:48 am local time
I build AI automation and CRM automation for B2B teams: Clay enrichment, n8n automation, lead qualification and scoring, AI research, outbound sequencing, CRM sync and reply routing.
I work with companies, agencies and growing startups that already have a sales process and a CRM in place, but need the infrastructure behind it to be more reliable, connected and scalable.
Before Upwork I spent 7+ years as a senior AI and automation engineer, including at two of Ukraine's largest banks, where a broken pipeline is not an inconvenience, it is an incident. MS in AI/ML. I write Python and SQL, not just no-code nodes, which matters once the logic outgrows what a visual builder can express.
Send me your current stack and where it breaks, and I will reply with how I would structure the fix. The architecture, not a sales call.
WHO THIS IS FOR
Established B2B companies, agencies and growing startups that already have a CRM, a defined ICP, and someone who owns sales. The motion exists. The system underneath it is manual, fragmented, or quietly dropping records.
I usually get called in when enrichment is half-manual and inconsistent, scoring lives in someone's head instead of in the pipeline, replies get triaged by hand, CRM data no longer matches reality, four GTM tools do not talk to each other, or the whole thing works at 200 records a month and falls over at 2,000.
WHAT I BUILD
Lead Enrichment & Data Infrastructure
Clay waterfalls with fallback logic and cost control. Custom API integration and scraper sources for the fields off-the-shelf providers miss. Verification and dedupe before spend, so records reach your CRM already clean.
Qualification & Scoring
ICP logic applied as a gate before enrichment spend, not as a report afterwards. Scoring rules expressed in code where the business logic matters, so they stay deterministic and auditable rather than a prompt returning a number nobody can reproduce. I have built prospect scoring systems as standalone data science projects as well as inside live pipelines.
n8n Automation & GTM Orchestration
n8n automation is the layer that holds the rest together: error handling, retries, idempotency, rate-limit backoff, dead-letter queues, alerting when a branch fails silently. This is the difference between a workflow that demos well and a system still running on a Tuesday when an API changes shape.
AI Research & Personalization
AI automation that reads a domain, job posts, filings or product pages and returns structured output your sequences can actually use, with schema validation and fallbacks so one bad model response does not ship a broken first line to 400 people.
Outbound Sequencing & Reply Automation
Smartlead and Instantly wired in properly: segment logic, suppression and exclusion rules, reply classification by intent, and CRM routing that sends each reply to the right owner with context attached.
CRM Automation & Routing
CRM automation for HubSpot and Pipedrive: lifecycle stages, deduplication, assignment and territory rules, enrichment on create, and two-way sync that does not fight itself. The CRM should reflect what is actually happening without someone repairing records by hand.
Reporting
Pipeline and campaign reporting built on the underlying records, so the numbers reconcile with your source data instead of with whatever a tool's native dashboard decides to show.
HOW I WORK
I start by mapping the current process and finding where records actually get lost. Usually it is data quality or routing. Rarely is it the sending tool. From there I build around the workflow you already have rather than forcing everything into a new tool. I instrument every workflow automation so failures are visible instead of silent, and hand over documentation so your team owns it. I take retainers on systems I have built, because AI automation fails quietly and someone should be watching it.
SCOPE
I build and own the technical system. I do not run daily campaign operations, do manual list building, or work on a pay-per-lead or pay-per-meeting basis. I can stand behind the engineering. Pipeline depends on your offer, your market and your sales team, and I would rather be straight with you about that than promise numbers I do not control.
STACK
Clay · n8n · Apollo · Smartlead · Instantly · HubSpot · Pipedrive · Apify · Python · JavaScript · SQL · OpenAI · Claude · RAG · Zapier · Make · Airtable · Supabase · REST APIs · webhooks
Tell me what is manual right now and I will tell you how I would build it, or whether it is worth building at all.
Steps for completing your project
After purchasing the project, send requirements so Bohdan can start the project.
Delivery time starts when Bohdan receives requirements from you.
Bohdan works on your project following the steps below.
Revisions may occur after the delivery date.
Initial Consultation
Discuss business needs, goals, and existing infrastructure to define the project's scope and expectations. Gather all necessary details and datasets for analysis.
Data Analysis & Preparation
Assess dataset quality, clean data, and prepare features to align with business objectives. Ensure data is ready for building a recommendation model.