About Me
AI/ML Engineer with 3+ years of experience developing and deploying machine learning, Generative AI, NLP, and computer vision solutions. Skilled in building LLM applications, RAG pipelines, AI agents, scalable inference …
AI/ML Engineer with 3+ years of experience developing and deploying machine learning, Generative AI, NLP, and computer vision solutions. Skilled in building LLM applications, RAG pipelines, AI agents, scalable inference APIs, and cloud-based ML services using Python, PyTorch, TensorFlow, Hugging Face, LangChain, FastAPI, Docker, MLflow, AWS, and GCP. Experienced across the complete ML lifecycle, including data preparation, model development, fine-tuning, validation, deployment, monitoring, and performance optimisation. Comfortable working across data, backend, frontend, and cloud environments within international, multidisciplinary engineering teams.
Experience
Software Developer - AI/ML
<p>Developed AI/ML, Generative AI, NLP, and computer vision solutions using Python, PyTorch, TensorFlow, Hugging Face, OpenAI, LLaMA, and cloud infrastructure.<br>
Built LLM applications, RAG pipelines, AI-agent workflows, conversational AI systems, and vector-search solutions using LangChain, Pinecone, FAISS, and ChromaDB.<br>
Fine-tuned transformer and LLaMA models with LoRA and PyTorch for domain-specific NLP and text-classification use cases.<br>
Developed and deployed ML APIs and inference services using FastAPI, Docker, AWS, GCP, and MLflow.<br>
Implemented data preparation, EDA, feature engineering, model validation, anomaly detection, OCR, ANPR, and model performance testing.<br>
Integrated AI services with frontend applications, enterprise systems, analytics platforms, and external tools.<br>
Reduced LLM response latency by 20% by integrating OpenAI models with Pinecone and AWS Lambda.<br>
Reduced average customer-support call duration by 25% through real-time voice AI agents.<br>
Accelerated reporting by 40% and reduced manual effort by 30% by developing Zoho Analytics pipelines.<br>
Increased user engagement by 15% by connecting backend ML APIs with frontend features.<br>
Developed computer vision and traffic-monitoring solutions using YOLOv5, OpenCV, PyTorch, PaddleOCR, and ANPR.<br>
Optimised inference performance through Docker deployment, asynchronous API workers, Locust load testing, and cloud-based scaling.</p>