Lead AI Engineer with 8+ years of experience leading teams and delivering production AI platforms from architecture through deployment. Builds Agentic AI, RAG, and generative AI systems that automate engineering workflow…
Lead AI Engineer with 8+ years of experience leading teams and delivering production AI platforms from architecture through deployment. Builds Agentic AI, RAG, and generative AI systems that automate engineering workflows. Has hands-on experience with multi-agent systems, LLM applications, computer vision, NLP, and speech recognition and voice AI, with a focus on translating business needs into reliable products.
<p>Led a team of engineers to design and deliver an autonomous documentation management platform embedded in the Git merge workflow, automating code analysis, documentation synthesis, review, and Wiki publication.</p><p> Designed the agent orchestration architecture using LangGraph, specialized agents, persistent memory, and vector databases for context-aware, autonomous execution.</p><p> Onboarded 100+ repositories, processing millions of lines of code and keeping system documentation synchronized with code changes.</p><p> Built real-time production observability across the multi-agent platform, providing visibility into agent execution, workflow failures, and documentation quality.</p><p> Improved Tracy's enterprise Q&A by refining retrieval pipelines, chunking strategies, and embedding selection to reduce hallucinations and improve response quality.</p><p> Built hybrid retrieval and contextual ranking into the RAG pipeline, helping users find relevant information across enterprise documents.</p>
<p>Developed ML pipeline for document OCR, classification, and entity extraction to automate national ID validation (PAN, Aadhaar, etc.).<br>
Implemented MobileNet-based classifier, improving accuracy by 20% and inference latency by 28%.<br>
Built an in-house OCR engine using Tesseract and custom preprocessing to enhance data extraction accuracy.</p>
<p>Designed and improved in-house OCR validation tools for automotive dashboards using image processing and multilingual text detection.<br>
Led feature implementation and performance optimisation for internal toolkits used across multiple teams.</p>
<p>Generated G-Codes for portrait drawing via image processing and API integration with DOBOT.<br>
Designed stereo-vision tracking and Kalman filter-based trajectory prediction for shuttlecock movement.</p>
<p>Architected and deployed an AI-powered video generation platform that automated end-to-end content creation workflows, significantly reducing manual effort and accelerating content delivery timelines.</p><p> Designed an LLM-driven workflow that autonomously analyzed narrative inputs, extracted scene structures, and mapped relevant visual elements to generate context-aware videos.</p><p> Engineered an intelligent metadata and asset retrieval framework enabling semantic understanding and context-aware orchestration of media components.</p><p> Optimized prompting strategies and generation workflows, improving output quality by 42% based on internal QA evaluation metrics.</p><p> Extended the existing platform with speech-to-video capabilities, enabling users to generate videos directly from voice inputs.</p><p> Designed and implemented an ASR pipeline integrating speech transcription and downstream content generation workflows for seamless voice-driven video creation.</p><p> Evaluated DeepSpeech, Wav2Vec, and Whisper through comparative POCs; deployed in-house Whisper-based solution, reducing operational costs by 28%, lowering latency, and improving transcription accuracy.</p>
<p>Designed and implemented code for angular velocity estimation using efficient image processing techniques.<br>
Integrated Raspberry Pi and Pi Camera for live portable testing; results matched expected theoretical values.</p>
نظام التوظيف التشغيلي المعتمد على الذكاء الاصطناعي أولاً، الذي يوحّد ويؤتمت دورة التوظيف بالكامل للمؤسسات الحديثة حول العالم.
د.جوب منطقة حرة ذ.م.م. 2026 © جميع الحقوق محفوظة
