AsifN

AI Machine Learning Engineer

F & A Global • Full time

Jan 2025 - Feb 2026 • 1 yr 1 mo

Contributed to the development of an AI-powered carbon emissions measurement and reporting system, automating end-to-end sustainability workflows and delivering measurable impact across data accuracy, processing speed, and operational efficiency. Key Contributions & Technologies: Carbon Emissions AI Automation: Built intelligent automation pipelines for carbon emissions measurement and ESG reporting, reducing manual effort by 80% and accelerating reporting cycles by 60% — enabling organizations to meet sustainability compliance requirements faster. Enhanced AI SQL Agents: Upgraded and extended AI-powered SQL agents for automated data extraction, validation, and transformation across large-scale sustainability datasets, achieving 25% improvement in data accuracy and 40% reduction in processing time. LLM Fine-Tuning (Llama 3.1 8B): Processed and curated 10,000+ domain-specific records to fine-tune Meta Llama 3.1 (8B) for healthcare and sustainability use cases. Implemented adapter-based fine-tuning using PEFT (Parameter-Efficient Fine-Tuning) with instruction tuning, boosting task-specific model performance by 30%. FastAPI Backend & Batch Processing: Developed scalable RESTful backend APIs with FastAPI and engineered batch processing pipelines for large dataset ingestion and transformation, increasing overall workflow efficiency by 50%. Cross-Functional AI Solutioning: Collaborated with cross-functional teams to translate complex healthcare and sustainability domain requirements into production-ready, scalable AI solutions Accelerating project delivery by 30%.

AI Engineer

Visnext Software Solutions • Full time

Dec 2023 - Dec 2024 • 1 yr

Architected and deployed a production-grade Conversational AI system for IoT smart home environments, enabling real-time natural language querying and intelligent monitoring of 10,000+ sensor and device records. Leveraged state-of-the-art Large Language Model (LLM) technology to translate user queries into precise, context-aware database operations. Key Contributions & Technologies: AI-Powered SQL Agent Development: Built intelligent SQL agents by integrating TimescaleDB with PostgreSQL (psql), achieving a 40% improvement in time-series query efficiency — ideal for high-frequency IoT sensor data workloads. End-to-End LLM Pipeline Engineering: Designed and optimized full LLM inference pipelines using DeepSeek R1 LLaMA 70B (distilled model), fine-tuned for IoT sensor data ingestion, real-time inference, and low-latency response generation. Scalable Backend Architecture: Engineered a robust, high-performance backend using FastAPI and PostgreSQL + TimescaleDB, supporting concurrent smart home device streams with RESTful API endpoints and efficient time-series data management. Natural Language to SQL (NL2SQL): Implemented NL2SQL capabilities allowing non-technical users to interact with complex IoT databases