Deepa Pandey — Software Engineer

Currently working on Generative AI, contributing to optimization, preparation, conversion, and execution of multiple foundation models (including LLaMA variants and others). Work involves model optimization, format conversion, quantization, and enabling efficient on-device inference workflows across Android and Snapdragon-based targets.Previously worked on the QAIRT Development Python API, migrating functionality from existing QAIRT CLI tools into structured, Python-native workflows. Built APIs covering model conversion, quantization, ahead-of-time compilation, execution, and analysis, enabling end-to-end execution of ML and GenAI models without direct CLI usage. Contributed to support for ONNX, PyTorch, TFLite, and GGUF models, execution across CPU, GPU, HTP, DSP, and AIC backends, and clear separation between host-side model preparation and target-side runtime execution across Linux, Windows, Android, and Snapdragon devices.Earlier, worked on Qualcomm AI Studio, building backend systems in Python using FastAPI. Developed REST APIs and real-time execution workflows using Redis, Celery, message queues, and WebSockets. Contributed to systems that run, convert, optimize, and quantize models (including ONNX), and prepare them for on-device deployment across Android, Windows, Linux, x86, and multiple hardware targets, in a fully Dockerized environment integrated with internal CI/CD pipelines.

Stackforce AI infers this person is a Backend-heavy Fullstack Engineer specializing in AI and machine learning technologies.

Location: Bengaluru, Karnataka, India

Experience: 4 yrs 7 mos