I bridge the gap between AI prototypes and production-grade systems. Specializing in high-performance FastAPI backends, scalable RAG pipelines, and fine-tuned LLMs engineered for global scale.
Top 3% Vetted Toptal Engineer | Available for Full-Time Remote Roles (US / UK / EU Timezones) | Contract-Ready via Deel
What I Offer
From concept to deployment, I provide end-to-end development solutions tailored to your business needs.
Building context-aware applications using LangChain, OpenAI/Gemini APIs, and advanced semantic chunking.
Designing robust, asynchronous RESTful APIs using FastAPI and Flask, optimized for low latency.
Structuring web scrapers and automation pipelines using Playwright and Selenium to feed production AI data streams.
Scaling machine learning workloads using Ray clusters and AWS EC2 autoscaling groups.
Ensuring identical development and production environments with Docker, Nginx, and automated CI/CD workflows.
A comprehensive overview of my technical skills and areas of expertise
My Work
Technical deep-dives into production-grade AI systems, datasets, and pipelines built for international clients.
Contributed to Google's Gemini development for 1.2 years, engineering SFT datasets for complex structured data reasoning and performing RLHF evaluation on automated model outputs.
Built a semi-supervised ML pipeline (SVM + CRF + DFS tree traversal) that classifies text and extracts data from German tender PDFs with 95% production accuracy.
Independently built a full-stack document extraction engine using Google Cloud Document AI and OCR to automate tax form and contract data extraction.
My Journey
Over 5 years of professional experience building digital products.
Toptal
Accepted into Toptal's top 3% network of global freelance engineers, vetted through a rigorous multi-stage technical screening. Available for senior full-stack and AI integration engagements through the network.
Scanify Dev
Contributed to an AI-powered Shopify theme generation platform that used OpenAI APIs to automatically create store content and images for ecommerce clients.
Turing.com (Google Gemini Project)
Contributed to a Google-commissioned project through Turing to improve Gemini's data analysis reasoning capabilities.
Recursive Data Ltd.
Contributed to multiple short-term projects spanning web scraping, browser automation, and AI integration.
Zaytrics (Pvt.) Limited
Independently built and deployed multiple production AI systems as client-facing APIs.
How I Work
Hiring globally shouldn't mean managing dependencies across time zones. I've designed my workflow to blend seamlessly into Western product teams.
Contract-ready immediately via Deel or B2B corporate billing. You don't deal with international payroll complexities.
Fully compliant with US/EU intellectual property laws under standard international contractor agreements.
Operating from a dedicated home office with secondary backup power (UPS/Solar) and high-speed fiber internet to ensure 100% uptime.
I utilize Slack, Loom video handovers, and clear Jira documentation so work moves forward without needing constant live synchronization.
I offer 4 to 5 hours of structured, synchronous daily overlap with EST, GMT, and CET business hours.
Seamless async-first engineering updates, Slack/Loom check-ins, and direct alignment with technical stakeholders.
Questions?
Find answers to common questions about my expertise and process.
I work PKT (UTC+5), which gives me natural overlap with US East Coast mornings (my afternoon) and full coverage of EU business hours. I use async-first workflows — daily written updates, Loom videos for demos, and a shared project board so you always know where things stand without needing a call. For time-sensitive projects, I'm flexible on hours.
Both. I'm open to full-time remote positions and long-term contracts. I'm set up to work as a remote contractor through Deel, or direct contracts. If you're hiring for a senior AI/backend role and want someone fully embedded in your team, let's talk.
I specialize in building production-ready Python backend architectures and AI systems. My day-to-day stack includes FastAPI, Flask, AWS, Docker, LangChain, and OpenAI/Gemini integrations. I also build robust frontends using ReactJS when a full-stack solution is required.
Absolutely! I offer flexible engagement models including full project ownership, team augmentation, and technical consulting. I can integrate with your existing workflows using tools like Jira, Slack, and GitHub. I'm comfortable working in both independent and collaborative environments.
Yes, I understand the importance of protecting your intellectual property. I'm happy to sign NDAs before discussing project details. All project code and data are kept strictly confidential and are never shared with third parties.
Post-launch changes are handled through our support agreement. Bug fixes are covered during the free support period. For feature additions, I provide a separate quote. I also offer monthly retainer packages for ongoing development needs.
Yes! I specialize in integrating AI capabilities into existing applications. This includes adding chatbots, recommendation systems, predictive analytics, image recognition, natural language processing, and more. I can assess your current system and recommend the best AI solutions for your needs.
Still have questions?
Let's TalkClient feedback
International clients across the US, Panama, and South Asia — several returning for multiple projects.
Research contributions to the field of Natural Language Processing and Machine Learning
Investigated and analyzed various text similarity metrics to improve the quality of abstractive summarization on the CNN/DailyMail dataset. The research focused on enhancing the coherence and relevance of generated summaries through advanced similarity measurement techniques.
This research presents a comprehensive comparative analysis of various text similarity metrics and their impact on abstractive summarization quality. Using the CNN/DailyMail corpus, we evaluated multiple similarity measures including semantic, syntactic, and hybrid approaches to determine their effectiveness in improving summary generation. Our findings demonstrate significant improvements in summary coherence and factual accuracy when incorporating optimized similarity metrics into the summarization pipeline.