S
Sanskar Unkule
About
Detail
Maharashtra, India
Résumé
Jobs
verified_user
0% verified
Prompt EngineerOutlier.aiNov 2024 - Current (1 year 9 months)• Collaborated as a freelance Prompt Engineer at Outlier on diverse projects, including Cypher RLHF Safety, Multilingual Centralized Screening, Kepler RLHF, and Association Plowman.. • Evaluated and provided detailed feedback on model-generated responses by employing Reinforcement Learning with Human Feedback (RLHF) techniques. • Assessed prompts and model outputs to identify performance gaps and refine instruction sets, ensuring improved model accuracy and reliability. • Contributed to fine-tuning the model through iterative feedback, enhancing its performance and alignment with project objectives.- ZPrompt EngineerZensar Technologies,Oct 2024 - Current (1 year 10 months)• Developed and refined prompt structures to improve AI model accuracy and reliability, focusing on advanced language model training and optimization under NVIDIA's NeMo project. • Collaborated with cross-functional teams to customize prompts for varied applications, ensuring alignment with NVIDIA's project objectives and maintaining high standards of output quality. • Generated dense captions for video data and contributed to chatbot creation planning, enhancing the project's data annotation and quality processes.
Education
verified_user
0% verified
- Bachelor of EngineeringSIES Graduate School of TechnologyJan 2020 - Jan 2024 (4 years 1 month)
Projects (professional or personal)
verified_user
0% verified
- OCypher RLHF SafetyOutlierAINov 2024 - Current (1 year 9 months)• Applied Reinforcement Learning with Human Feedback (RLHF) to evaluate and provide feedback on user-entered prompts and model-generated responses. • Identified improvement areas in prompt-response interactions to enhance model performance and alignment with desired outcomes. • Contributed to the iterative refinement of the model by ensuring high-quality feedback and effective instruction adjustments.
- Dense Video Captioning Commercial VITAZensarOct 2024 - Current (1 year 10 months)• Annotated video datasets by crafting precise event captions and global captions to facilitate accurate model training for dense video captioning. • Performed Quality Assurance (QA) to ensure high-quality annotations, validating data accuracy and consistency to optimize model performance. • Contributed to enhancing the efficiency of AI models by delivering reliable and comprehensive video captioning datasets..
- IInformation Retrieval Using Evolutionary AlgorithmJul 2023 - Apr 2024 (10 months)• Developed a comprehensive system for efficient information retrieval using evolutionary algorithms. • Implemented inverted indexing and an advanced document indexing method. • Applied modified genetic algorithm, cultural algorithm, ant colony optimization, particle swarm optimization, and bee swarm optimization to enhance search accuracy. • Demonstrated proficiency in information retrieval techniques and evolutionary computing.