Aditya Srinivas Timmaraju
About Me
(Updated as of 07/10/2017) I recently joined Facebook as a Software Engineer.

I’m a Research Scientist at Blippar, an AR/AI startup, working on deep learning applied to computer vision. I graduated in June 2015 from Stanford with a Masters in EE. I earned my Bachelor of Technology (Honors) with a major in EE and minor in CS at the Indian Institute of Technology (IIT) Hyderabad, where I won the Gold medal.

I am experienced in applying Deep Learning for Visual Recognition tasks such as fine-grained attribute recognition, human pose estimation, object detection, text recognition and for NLP tasks such as Sentiment Analysis, Named Entity Recognition and Language Modeling.
Interests
Machine Learning, Computer Vision, Natural Language Processing
Contact
You can reach me at tadityasrinivas@gmail.com
Previous
CS 224D — Sentiment Analysis on Movie Reviews using Recursive and Recurrent Neural Network Architectures
Proposed a novel architecture that uses a Recursive NN within a sentence, and a Recurrent NN over the hidden states of the Recursive NN, to process paragraphs such as movie reviews. This processing mechanism resembles the way humans read. (Tools: Python, wrote RNNs, RecNNs from scratch)
CS 231N — Detecting and Recognizing Text in Natural Images using CNNs
Built a system for end-to-end text recognition in natural images, using two different CNNs, one for centered Character Detection and another for Character Recognition. (Tools: Python, Caffe)
CS231A — Learning Binary Descriptors from Images
Designed a novel compact machine-learned binary descriptor that outperforms FREAK on Patch dataset
Independent Study — A Maximum Likelihood Approach to Localization from Connectivity Information
Used a maximum-likelihood approach to solve the localization problem. Characterized the Fisher Information Matrix of the problem.
Sole TA for CS379C — Computational Models of the Neocortex
Spring 2015, taught by Prof. Tom Dean, former Deputy Provost and chair of CS, Brown University. I helped students with setup, installation and usage of deep learning frameworks viz., Caffe, MatConvNet. Projects in the class included structural and functional connectomics of neuronal images.
TA for CS231A — Computer Vision: From Recognition to 3D Reconstruction
Winter 2015, taught by Prof. Silvio Savarese. I was responsible for authoring some chapters of the currently used lecture notes.
TA for EE263 — Linear Dynamical Systems
Fall 2014, taught by Prof. Sanjay Lall. Created new problems for the mid-term and final exams. Conducted tutorial sessions and office hours.
Graduate Research — Computational Vision and Geometry Lab
Iterative clustering to discover discriminative mid-level patches in chair images.
Qualcomm Research Silicon Valley — Santa Clara, CA
Software Engineering Intern. Developed a new particle filtering based SLAM algorithm for indoor mobile robots. (Tools: ROS, C++, Python)
Qualcomm Inc — Bangalore, India
Undergrad hardware internship. Worked on RTL, functional verification. (Tools: Verilog)