[email protected] | LinkedIn | craftsmanlabs.net
GitHub | Chennai, India | +91 98849 88197
AI/ML Engineer with deep expertise in LLMs, GenAI, and enterprise AI solutions. Specializes in architecting cost-optimized, production-ready AI systems with a focus on regulated industries and synthetic data generation. Adept at leveraging cross-domain research to drive innovation and operational excellence.
Founder & AI Consultant | Sept 2024 – Present
- Shipped production-ready REST APIs using FastAPI with async support, rate limiting, and OAuth2 authentication for enterprise LLM services
- Architected AI-as-a-Service solutions leveraging DeepSeek-7B-R1, GPT-4, and Gemini Pro for enterprise clients
- Developed agentic RAG systems using PineCone, Azure AI Search, and LiteLLM, optimizing response generation with Groq inference
- Built scalable ML pipelines using SageMaker and Azure ML Studio, integrating queues for high-throughput processing of client workloads
- Implemented vision-language solutions using QWen-VL 2.5 and DeepSeek for multimodal understanding in production environments
- Specialized in fine-tuning open-source LLMs (7B-70B parameter models) for domain-specific applications and cost optimization
- Delivered end-to-end AI solutions for financial, healthcare, and enterprise clients, achieving 90% accuracy in compliance verification
- Leveraged cross-domain insights to build custom LLM applications integrating medical, financial, and operational workflows
Software Engineer ML | July 2023 – Sept 2024
- Built multimodal LLM solutions (Donut, GPT-4v, Llava) for Intelligent Document Processing, enhancing data extraction and processing
- Developed automation sales bots using custom LLMs, streamlining lead conversations and cutting operational time by 70–80%
- Spearheaded a web scraping initiative identifying 500+ distressed assets in China to drive actionable supply chain insights
- Integrated AI tools with Gmail, HubSpot, Telegram, and Slack for improved operational efficiency
Solutions Engineer | Nov 2022 – June 2023
- Accelerated financial document data extraction by 100% using LayoutLMv3, achieving 85% accuracy on handwritten forms
- Integrated Intelligent Document Processing (IDP) pipelines with proprietary bank systems to streamline document workflows
- Collaborated with global teams to refine and optimize data extraction methodologies
Software Engineer ML | May 2022 – June 2023
- Developed a signature validation system using RNNs and YOLOv5, attaining 85% accuracy
- Built an invoice analyzer leveraging LayoutLMV2/V3 for multilingual document processing
- Implemented an MLOps pipeline incorporating human-in-loop annotation for continuous model improvement
- LLMs & GenAI: GPT-4o, Gemini Models, Qwen Models, OpenELM models, Vision Lnaguage models like Janus Pro, Mistral, LLaVA, Custom Fine-Tuning, RAG, Synthetic Data Generation
- ML Frameworks: PyTorch, TensorFlow, Langchain, LlamaIndex
- Computer Vision & NLP: YOLO, LayoutLM (v1/v3), Transformers, BERT, Word Embeddings
- Languages: Python (FastAPI, Django, Flask), C++, CUDA, JavaScript (React)
- Cloud & DevOps: AWS (EC2, Lambda, DynamoDB), Azure, Docker
- Databases: SQL, DynamoDB, Redis, Vector Databases
BS Abdur Rahman Crescent Institute | 2018–2022 | CGPA: 9.6
- Class Top Percentile Student (2022) – CGPA: 9.6
- Multiple coding competition wins at Crescent Coding Club
- Finalist in various hackathons including SIH