Barada Acharya

AI Researcher

Santa Clara, California, United States10 yrs 6 mos experience

Most Likely To SwitchHighly Stable

Key Highlights

Led development of state-of-the-art NLP models at Apple.
Pioneered multilingual NLP capabilities for iOS/macOS.
Improved user typing experience for billions with advanced models.

Stackforce AI infers this person is a Machine Learning expert specializing in Natural Language Processing within the SaaS industry.

Contact

baradaraja@gmail.com LinkedIn

Skills

Core Skills

Natural Language Processing (nlp)Machine Learning

Other Skills

ASP.NETAlgorithmsCC#C++Core JavaData StructuresDeep LearningJavaLarge Language Models (LLM)LinuxMatlabMultilingual NLPOracle SQLProgramming

About

I am an ML Research Engineer at Apple in the field of NLP. As an ML Research Engineer, build and ship state-of-the-art NLP models and frameworks for Apple's NLP team(Apple Intelligence) part of Intelligence system experience, SWE.

Experience

10 yrs 6 mos

Total Experience

3 yrs 6 mos

Average Tenure

7 yrs 7 mos

Current Experience

Apple

2 roles

Senior ML Research Engineer

Promoted

Oct 2020 – Present · 5 yrs 8 mos

As a Modeling DRI,
2023-24:
Led the design and modeling of the "Describe your change", a prompt-based text-editing feature in Apple Intelligence Writing Tools, facilitating direct user interaction with Apple Foundation Model (AFM). This involved formulating problem statements, preparing data, defining eval strategy, SFT and RLHF training on AFM.
Designed initial experiments and eval pipeline to generate relevant visual prompts from text paragraph for the Stable Diffusion model using AFM, facilitating the creation of images from Notes (Image Wand).
2022-23:
Led the development of a multilingual transformer-based Language Model for 20 Latin languages, facilitating on-device application development across iOS/macOS (e.g. a multilingual stable diffusion model). This milestone set the stage for future language expansions and rollout of several multilingual iOS keyboard experiences.
2020-22:
Delivered an English transformer-based Language Model for iOS/macOS keyboards. This model revolutionized iOS typing experience by vastly outperforming previous LSTM-based models, driving autocorrection, proofreading, in-line prediction and next-word prediction typing features for billions of users.

Machine LearningNatural Language Processing (NLP)Multilingual NLPLarge Language Models (LLM)Deep Learning

ML Research Engineer

Oct 2018 – Sep 2020 · 1 yr 11 mos

Implemented the initial versions of Transformer and BERT models on macOS/iOS, marking an important milestone for bringing Transformer technology on-device and setting the stage for future advancements and shipments.
As the Modeling DRI, I delivered key NLP technology, including transfer learning for text, which is widely used by first-party and third-party developers. This effort also involved shipping static word embeddings, ELMo-based contextual embeddings, and sentence embeddings for multiple Latin languages.

Amazon

2 roles

Applied Scientist Intern

May 2018 – Aug 2018 · 3 mos · Greater Seattle Area

Developed a new objective function for content optimization in Amazon's recommendation service using Reinforcement Learning to enhance reward quantification and reduce bias and variance.
Ensured alignment with online A/B testing and conducted extensive data analysis and visualization with terabytes of data.

SDE Intern (Machine Learning)

Jun 2017 – Sep 2017 · 3 mos · Greater Seattle Area

I developed a supervised classifier to resolve conflicts in brand classification for Amazon's Selection and Catalog Systems. By creating a weighted matching algorithm that converted text and web content into a vector space model, we reduced manual auditing and minimized classification errors.

Uc irvine

3 roles

NLP Research

Promoted

Sep 2017 – Mar 2018 · 6 mos

Worked with Dr. Sameer Singh on knowledge extraction from diverse unstructured data (image, video, text). Experimented with text summarization using seq2seq architecture (LSTM/GRU with attention) for description generation from multimodal data.
Responsibilities:
Reviewing literature and understanding past work. Proposing model design. Implementing and analyzing models (Python, Keras, PyTorch, TensorFlow).

Reader

Apr 2017 – Mar 2018 · 11 mos

Reader for the course (boolean algebra and logic) and (Scientific Computing).

Bayesian Learning Research

Apr 2017 – Mar 2018 · 11 mos

I have worked with Dr. Padhraic Smyth on Bayesian deep learning using Variational inference and Laplace approximations and investigated on VAE based Deep Generative models. I have implemented a novel method for a deep generative model by using adaptive recurrent inference networks and analysed its results using Python and TensorFlow.

Samsung electronics

Software Engineer

Jul 2014 – Jul 2016 · 2 yrs · Banglore

Implemented multi-device support for Memory Solution Tool Kit (MSTK)
Designed the NVMe-compatible core library for SSD Verification Platform (SVP), supporting Windows and Linux.
Integrated CINT and Python with SVP script executor, ensuring compatibility with SVP core library.
Received a spot award for MSTK; the project won Outstanding Project of the Year and the President Award.

Defence research and development organisation (drdo)

Summer Intern

May 2013 – Jul 2013 · 2 mos · Chandipur, Odisha

Study and implementation of Zoom Tracking and Auto-focus methods on Video Cameras.
Implemented and compared various focus-region selection methods, sharpness measurement functions and auto-peak-search algorithms for static and moving objects. Subsequently, developed a software using Relational Zoom Tracking algorithm enabling focus and zoom control methods based on the distance of the target objects. Used MFC.