Speech Technologies are breaking down language barriers by fostering clearer, effective and easier communication across diverse global communities.

Priti Raychoudhury (PhD), our next pathbreaker, works as a Senior Linguist at Sanas, an AI startup specializing in real-time accent translation technologies.

Priti talks to Shyam Krishnamurthy from The Interview Portal about discovering her fascination not with literature but with the analytical aspects of language and the captivating field of computational linguistics.

For students, the landscape of technology and linguistics is ever-evolving. There is a lot to explore, learn and build !

Priti,  Your background?

I’m an Army kid, which meant my childhood was a bit of a grand tour around the northeastern parts of India. A good chunk of my early years was split between the quaint town of Silchar in Assam and the scenic beauty of Shillong in Meghalaya. I’ve always been fascinated by human language—its structure, its analytical aspects, and how we use it to connect with each other. This passion for languages, I guess, is something I picked up from my parents. They’ve always been keen on understanding and appreciating the diversity of languages around us, and that curiosity was instilled in me from a very young age.

What did you do for graduation/post graduation?

For my undergraduate degree, I pursued English, thinking my interest lay in literature. However, it didn’t take long for me to realize that my real passion was not in the stories themselves but in the language used to tell them. Eager to dive deeper, I went on to complete my Master’s in Linguistics at NEHU (North Eastern Hill University), where I fully embraced the analytical side of language, setting the stage for my future endeavors in the field. This academic journey set the stage for my subsequent work with the Mother Tongue Survey of India and at Google Bangalore, before I embarked on my PhD in Linguistics, specializing in experimental phonology and acoustic phonetics.

What were some of the drivers that influenced me to pursue such an offbeat, unconventional and uncommon career in Linguistics?

The turning point was a linguistics course I took during my major. It was like a lightbulb moment, showing me that language could be analyzed and understood from a scientific perspective. This realization was a game-changer for me. Coming from a small town, I wasn’t initially aware of the vast opportunities in linguistics beyond academia. 

Choosing this career path in linguistics, especially focusing on experimental phonology and acoustic phonetics within the realm of speech technology, was influenced by a confluence of personal inclinations, influential individuals, pivotal events, and a series of turning points that collectively steered me towards my current role.

a) Key Influencers: My upbringing as an Army kid, constantly moving across diverse linguistic landscapes of northeastern India, notably Silchar in Assam and Shillong in Meghalaya, played a significant role. This exposure to a rich tapestry of languages ignited my curiosity about human language and its complexities from a very young age.

b) People/Mentor: A pivotal figure in my academic and professional journey was Dr. Umarani Pappuswamy, my lecturer in computational linguistics at NEHU. Dr. Pappuswamy, with her inspiring career across various countries, introduced me to the potential of applying linguistic principles in the industry. Her encouragement and guidance were instrumental in my decision to delve deeper into applied linguistics.

c) Events: My initial exposure to linguistics during my undergraduate studies, where I discovered my fascination not with literature but with the analytical aspects of language, marked a critical event in my career trajectory. This realization propelled me to pursue a Master’s in Linguistics at NEHU, where I encountered Dr. Pappuswamy and the captivating field of computational linguistics.

d) Turning Points: Working as a graduate researcher in the Mother Tongue Survey of India and as a speech data evaluator at Google Bangalore were significant turning points. These experiences exposed me to the practical applications of linguistics and the power of speech technology. My subsequent decision to pursue a PhD, focusing on the digitalization of North Eastern languages and delving into experimental phonology and acoustic phonetics, stemmed from a desire to explore how speech signals correlate with linguistic signals. The marvel of the human mind’s ability to encode and decode these details effortlessly has always fascinated me and motivated my research and career choices.

The culmination of these influences, mentors, events, and turning points not only shaped my career trajectory but also reinforced my commitment to advancing the field of speech technology through linguistics.

How did you plan the steps to get into the career you wanted? Or how did you make a transition to a new career? Tell us about your career path

Initially drawn to the analytical aspects of language, my career has been guided by a fascination with the nuances of human speech and its digital preservation. The desire to bridge theoretical linguistics with practical technological applications has been my driving force, aiming to make tangible contributions to speech technology.

My career path in linguistics and speech technology has been a blend of academic rigor, practical application, and continuous learning, shaped significantly by my experiences across diverse roles and institutions. Here’s a comprehensive look at how I navigated my journey from academia to the industry, highlighting key positions and educational milestones.

I started my career as a Graduate Research Assistant at the Ministry of Home Affairs, Government of India. My professional journey began with fieldwork focused on the linguistic documentation of two lesser-known languages of northeastern India. This role involved extensive phonetic transcription, phonological, morphological, syntactic, and semantic analysis, laying the groundwork for my interest in digital language preservation.

My next role was as a Speech Data Evaluator at Google, Bengaluru (Nov 2012 – Aug 2013): Moving into the tech industry, I specialized in phonetic transcription, Word Sense Disambiguation, and annotations for Indian English and Hindi. This experience deepened my understanding of the application of linguistics in technology, particularly in speech recognition.

My work at Google was part of a product that many are familiar with – the Google Voice search feature. During my tenure as a Speech Data Evaluator at Google Bengaluru from November 2012 to August 2013, I was involved in a Text-to-Speech (TTS) project that contributed to enabling voice search functionality on Google. This project aimed to enhance the accuracy and efficiency of voice recognition and synthesis for Indian English and Hindi languages, thereby improving the overall user experience of Google’s voice search feature.

After that, I was hired as a Linguist (Junior Project Fellow) at the Indian Institute of Technology, Guwahati (Jan 2014 – Aug 2017): Under the guidance of Prof. Shakuntala Mahanta, I worked on the digital preservation of twelve northeastern languages. This project not only honed my skills in acoustic phonetic and experimental phonological analysis but also instilled in me the importance of preserving linguistic diversity through technology.

With all these experiences, I decided to do a PhD in Linguistics from the Indian Institute of Technology, Guwahati (Jul 2014 – Dec 2022): My doctoral research, supervised by Prof. Shakuntala Mahanta, explored the prosodic structure and tonogenesis of Sylheti, marking a significant milestone in my academic career. This work underscored the intersection of phonology, technology, and linguistic diversity, propelling me towards a career in speech technology.

Sylheti is a language, spoken primarily in Barak Valley of Assam, Indian and the Sylhet region of Bangladesh, and also among the Sylheti diaspora around the world, notably in the United Kingdom.

The practical applications of my PhD research in understanding the prosodic structure and tonal aspects of Sylheti: imagine you’re using a voice recognition system, like the one you might find in your smartphone or smart home device. Now, consider a scenario where this system needs to accurately understand and interpret spoken commands or queries in Sylheti. 

My research has the potential to improve the precision of such systems through insights into the distinctive tonal features of Sylheti.

By investigating the phonological tone in Sylheti and understanding how it developed independently within its non-tonal language family, my research sheds light on the tonal nuances crucial for accurate speech recognition and synthesis. This means that through empirical acoustic evidence and tonogenetic analysis, we gain a deeper understanding of how tones function in Sylheti at various linguistic levels beyond just the words themselves. So, in essence, my PhD work has tangible applications in advancing speech technology, making it more adept at recognizing and responding to the linguistic diversity present in our world, including languages like Sylheti. 

The academic implication of my PhD research lies in advancing our understanding of tonogenesis and tonal interaction in languages like Sylheti, contributing to the broader field of phonological typology.

How did you get your first break?

I would attribute my major breaks to

Contacts and Networking: Building a professional network has been crucial throughout my career. From mentors like Prof. Shakuntala Mahanta and Dr. Umarani Pappuswamy to collaborations with a consortium of northeastern universities, these connections have provided invaluable support and opportunities for growth.

Continuous Learning and Adaptation: The dynamic nature of speech technology and linguistics demands ongoing learning. Attending workshops, seminars, and engaging in interdisciplinary research have been essential strategies for staying current and contributing innovatively to my field.

Each step of my career, from initial internships and research roles to my PhD and current industry position, has been carefully chosen to build upon my expertise in linguistics and speech technology. This journey reflects a deliberate effort to merge my academic interests with practical applications, aiming to contribute meaningfully to the advancement of speech technology and the preservation of linguistic diversity.

My initial leap into the field came with my role at Google, where I worked as a speech data evaluator. This position opened my eyes to the vast and intriguing world of Natural Language Processing (NLP), sparking a profound interest in the speech components of linguistics. It was this experience that motivated me to delve deeper into the specialized area of speech technology.

What were some of the challenges you faced? How did you address them?

Challenge 1: Venturing into an unconventional field was my first major challenge. Coming from a background where traditional careers like medicine and engineering were the norm, choosing linguistics and specifically, a niche within speech technology, was stepping well beyond my comfort zone. To overcome this, I embraced the responsibility for my choice, reassuring myself and my family that this path, though less traveled, was one I was passionate about and committed to. This mindset helped me to stand firm in my decision, without harboring any regrets.

Challenge 2: Transitioning from academia to the corporate world presented another significant hurdle. Moving to Bangalore for employment opportunities was a leap of faith, coupled with months of uncertainty and struggle. The turning point came when I received the opportunity to join Google as a speech data evaluator. This experience taught me the value of perseverance and adaptability, demonstrating that embracing change and facing challenges head-on can lead to rewarding opportunities.

Challenge 3: Securing a position in a PhD program that aligned with my interests in linguistics was my third major challenge. I aspired to be part of an institution where research in language technology was emphasized, which led me to IIT Guwahati. Joining the project on the digitalization of northeastern languages and pursuing a PhD in experimental phonology and acoustic phonetics required a great deal of effort and persistence. Overcoming this challenge involved not only proving my academic and research capabilities but also aligning my aspirations with the goals of the institution and my supervisor, Prof. Shakuntala Mahanta.

In each instance, addressing these challenges involved a combination of personal conviction, perseverance, and a willingness to embrace new opportunities. These experiences have been instrumental in shaping my career and guiding me towards achieving my professional goals.

Where do you work now? What problems do you solve?

I currently work as a Senior Linguist at Sanas, Bengaluru. Sanas is an AI startup specializing in real-time accent translation technologies. Our product aims to facilitate seamless communication by accurately translating accents in real-time. This technology allows for more effective cross-cultural communication, benefiting individuals and businesses with diverse linguistic backgrounds on a global scale.

As a linguistics subject matter expert, I lead projects on Root Cause Analysis (RCA) of internal model issues, leveraging both objective and linguistic metrics. My role involves cross-functional coordination, training team members on linguistic tools and methodologies, and enhancing model performance through acoustic phonetic, phonological, and linguistic data analysis.

Sanas is an AI startup with a global footprint. I apply my extensive background in linguistics research and practical expertise in speech data analysis to improve real-time accent translation technologies.

What skills are needed for your role? How did you acquire the skills?

My role involves conducting detailed evaluations of our models, pinpointing issues through root cause analysis, and offering actionable recommendations to refine and advance our translation systems. My work requires effective collaboration across different teams and regions, ensuring that our projects benefit from a wide range of expertise and perspectives.

I also hold a Master’s degree in Linguistics complemented by an understanding and hands-on experience in coding languages like python or R, and over 5 years of experience of working with speech data and linguistics of Indian languages from different language families.

My proficiency in English, along with a keen understanding of various accents, imparts me with strong communication skills, both written and verbal, paired with the ability to present complex information clearly.

I have a knack for organization and detail, alongside robust analytical abilities developed through more than 5 years of experience in research and development, with publications in leading journals.

I also have 2 years of experience in data processing, including data visualization and analysis, and experience working with diverse, cross-functional teams on a global scale.

What’s a typical day like?

My typical day involves a mix of technical analysis, project management, and collaboration. I start by reviewing model performance reports, then dive into objective evaluation to enhance analytics, and communicate with stakeholders for the projects. 

My day also involves meetings with cross-functional teams to discuss findings and strategies for improvement.I also spend time mentoring team members, sharing insights on linguistic tools and methodologies.

What do you love About this job?

What stands out for me in this role is the direct impact of my work on enhancing speech synthesis across different cultures and languages through advanced speech technology. The challenge of solving complex linguistic problems by bridging the gap between speech signals and linguistic signals and the opportunity to work with a talented, diverse team are incredibly fulfilling aspects of my job at Sanas.

How does your work benefit society?

At Sanas, my role as a Senior Linguist contributes to society by breaking down linguistic barriers and fostering clearer, more effective communication across diverse global communities. Through enhancing real-time accent translation models, my work supports inclusivity and understanding in communication, ensuring that individuals, regardless of their native language or accent, can interact seamlessly and without prejudice. This technological advancement aids in diminishing the communicative divide, promoting equal opportunities in professional and personal spheres for speakers of all languages and dialects. By improving these models, we’re not just advancing technology; we’re also cultivating a more interconnected and empathetic world where every voice is heard and valued.

Tell us an example of a specific memorable work you did that is very close to you!

One of the projects that holds a special place in my heart involved leading objective evaluations of synthesized speech at Sanas. Utilizing my expertise in phonology and acoustic phonetics, my job allowed me to delve deeply into the nuances of speech technology, critically assessing and improving the naturalness and intelligibility of synthesized voices. This work not only challenged me intellectually but also provided a platform to apply my doctoral research in a practical, impactful way. My contribution in enhancing the quality of synthesized speech models has been a significant milestone in my career, reflecting my commitment to bridging the gap between academic research and real-world applications in speech technology.

Your advice to students based on your experience?

Follow your passion with Purpose: Choose a field that not only interests you but also holds significance for you personally. Your passion will fuel your perseverance through challenges and motivate you to make meaningful contributions.

Embrace Interdisciplinary Learning: The future belongs to those who can bridge disciplines. Combine your core interests with skills from related fields, such as computer science, cognitive science, or data analysis, to enhance your versatility and value.

Seek Mentorship and Collaborate: Connect with mentors who inspire you and can guide your path. Collaboration with peers and professionals from diverse backgrounds will enrich your understanding and open up new avenues for exploration.

Engage in Hands-on Projects: Theoretical knowledge is crucial, but the application of that knowledge is what makes it valuable. Seek out internships, research projects, or create your own projects to apply what you’ve learned in real-world contexts.

Stay Curious and Keep Learning: The landscape of technology and linguistics is ever-evolving. Cultivate a habit of continuous learning to stay abreast of the latest developments and theories in your field.

Communicate Your Ideas Effectively: Whether it’s through academic papers, presentations, or blog posts, learn to communicate your ideas clearly and compellingly. Effective communication is key to sharing your insights and making an impact.

Be Resilient and Open to Change: You will encounter challenges and setbacks. Embrace them as opportunities for growth. Be open to changing your direction if you discover new interests or if the industry evolves.

Contribute to Society: Consider how your work can benefit society. Whether it’s through improving technology, preserving endangered languages, or enhancing communication, aim to make your work impactful beyond the academic and professional spheres.

Remember, your career is a journey, not a race. Take the time to explore, learn, and grow, both professionally and personally. Your unique path will be enriched by the challenges you overcome and the discoveries you make along the way.

Future Plans?

Looking ahead, my aspirations are rooted deeply in the advancement of speech technology and linguistic research. I aim to be at the forefront of integrating artificial intelligence with the nuances of human language, navigating the dynamic landscapes of both industry and academia as needed. My goal is to contribute meaningfully to the digitalization of human languages, ensuring that this endeavor not only preserves linguistic diversity but also enhances communication technologies globally.

Being actively involved in AI-driven projects, I envision developing solutions that bridge the gap between human linguistic capabilities and machine understanding. I am particularly interested in seeing the real-time application of my research and knowledge, transforming theoretical insights into practical tools that facilitate more natural and accessible human-machine interactions.