Jan 2025 – Jun 2025 · 5 mos · Bengaluru, Karnataka, India · On-site

• Working with the Digital Acceleration team on an AI agent that will
• automatically resolve merge from live build failure, and reduce the
• dev effort estimated to be 5-10 hrs per week per team and help
• maintain pipeline freshness. We are using the MCP-based
• architecture and the Claude Sonnet 3.7 model for inference.
• Wrote a code for PEFT-based GRPO fine-tuning of open-source LLM
• (GPT-2, codellama-7b-hf & Quen-2.5-7b).

May 2024 – Aug 2024 · 3 mos · Bengaluru, Karnataka, India · On-site

• Worked on Benchmarking video embedding models for Embedding Based Retrieval to achieve accurate search results, improve recommendation systems, and remove harmful content. Made video embedding pipelines using Ray That can be integrated seamlessly with PyTorch and TensorFlow models for batch inferencing that archives 30X speedup. Also Curated a Dataset that can be further used for experimentation.

Aug 2023 – Jan 2025 · 1 yr 5 mos · Hyderabad, Telangana, India

• Working on developing efficient training algorithms for spike response model of neuron. Working with Dr. Venkatkrishna Ramaswami, who is expert in computational neuroscience.

Jan 2023 – May 2023 · 4 mos · Hyderabad, Telangana, India

• Worked with Dr. Aruna Malapati on fault detection in wind turbines using reinforcement learning and using wavelet transform to convert sensor signals to images.

May 2023 – Jul 2023 · 2 mos · Pilani, Rajasthan, India · On-site

• Worked on a Multimodal deep-learning architecture for anomaly detection in videos
• from surveillance camera under Dr Sumeet Saurav (Senior Scientist CEERI).
• Using public database UCF-Crime. Learned to deploy 3D CNN models using
• the Keras and TensorFlow.

Aug 2022 – May 2023 · 9 mos · Hyderabad, Telangana, India