About sakshi
I'm a Principal Software Engineer with 7 years of experience building and evolving a production AI/ML product at Interra Systems — a media tech company specializing in automated captioning, subtitling, and broadcast quality tools. My core work spans NLP, ASR, and applied LLMs — fine-tuning Whisper models for dialect detection, building RAG-based chatbots, designing NLP-driven segmentation APIs, and developing agentic AI systems for automated code review. I've also expanded multilingual captioning support to 20 languages and improved transcription sync accuracy from 94% to 97%. I find the most interesting problems at the intersection of backend engineering and AI — where systems need to be not just accurate, but reliable and scalable in production. Open to roles in: NLP Engineering · Applied AI/ML · Backend Engineering with AI focus
Key Skills
Experience
Principal Engineer
CurrentInterra Systems
* Built a dialect detector by fine-tuning Whisper on limited domain-specific datasets to classify regional Spanish dialects, improving localization accuracy * Designed and deployed a RAG-based customer support chatbot using Python and vector search, reducing support ticket volume * Fine-tuned Whisper on domain-specific vocabulary using DeepSpeed, reducing post-editing effort on transcriptions * Optimized Faster Whisper and WhisperX internal parameters, improving transcription accuracy and system stability under load * Designed an NLP-driven auto-segmentation REST API for linguistically accurate caption generation aligned with broadcast standards * Introduced a unified Python unit testing framework, increasing automated test coverage and enabling faster CI/CD releases
Senior Software Engineer
Interra Systems
* Re-architected backend workflows using the Chain of Responsibility design pattern, improving modularity across microservices * Integrated speaker diarization and profanity filtering into captioning pipelines using NLP, improving subtitle usability for international markets * Refactored spaCy NLP module loading into a singleton class, reducing memory overhead and startup time * Mentored 3 junior developers through onboarding, pair programming, and code reviews * Authored internal white papers, technical documentation, and developer guides
Software Developer
Interra Systems
* Developed an NLP-based segmentation algorithm for caption breaks aligned with Netflix captioning guidelines. * Expanded language support from English to 20 languages, enabling global deployment * Improved media sync error detection from 94% to 97% using a dynamic-programming-based transcript alignment algorithm * Implemented automated quality checks — spell check, accuracy validation, profanity detection — using Python NLP libraries * Added export support for TTML and WebVTT caption formats
Education
Indian Institute Of Information Technology Kota
Bachelors
Interested in connecting with sakshi?
Sign up for NinjaHire to send a connection request.
Common Questions
What is sakshi's expertise?
sakshi specializes in Fine-Tuning Engineer, with expertise in agentic ai development, algorithm development, audio processing, backend development, c++.
Where is sakshi located?
sakshi is based in noida, uttar pradesh, india.
How much experience does sakshi have?
sakshi has 7+ years of professional experience.
How can I contact sakshi?
You can connect with sakshi through NinjaHire by signing up for a free account.