About
Activity
-
TPM setting up fancy JIRA board, sprints, story point, backlog and burndown chart using 3 different tools for program V/S a real TPM executing…
TPM setting up fancy JIRA board, sprints, story point, backlog and burndown chart using 3 different tools for program V/S a real TPM executing…
Liked by Aman Gupta
-
PyTorch Distributed Shampoo Wins the External Tuning Track of the Inaugural AlgoPerf: Training Algorithms Benchmark! "The external tuning ruleset…
PyTorch Distributed Shampoo Wins the External Tuning Track of the Inaugural AlgoPerf: Training Algorithms Benchmark! "The external tuning ruleset…
Liked by Aman Gupta
-
Excited to share that I will be presenting our ( Ashvini Jindal, Ankur Parikh and me) work titled "DistFin: Distillation based Fine-Tuning for…
Excited to share that I will be presenting our ( Ashvini Jindal, Ankur Parikh and me) work titled "DistFin: Distillation based Fine-Tuning for…
Liked by Aman Gupta
Experience & Education
Patents
Courses
-
Algorithms for Natural Language Processing (PhD level course)
11-711
-
Control Systems
-
-
Data Mining
-
-
Data Structures and Algorithms
-
-
Database Systems
-
-
Introduction to Machine Learning (PhD level course)
10-701
-
Linear Algebra
-
-
Machine Learning for Large Datasets
10-605
-
Network Programming
-
-
Numerical Analysis
-
-
Operating Systems
Course Topper
-
Operations Research
-
-
Optimization
-
-
Pattern Recognition
-
-
Probabilistic Graphical Models (PhD level course)
10-708
-
Probability and Statistics
-
-
Programming Languages and Compiler Construction
-
Honors & Awards
-
Graduate Research Fellowship
Carnegie Mellon University
Tuition and stipend support
-
Winner at Customer Service Hackathon
Amazon
My team and I designed a scalable solution to enable customers to have a better experience at Amazon.com.
-
AIMA - All India Merit Scholarship - AIR 2
AIMA
Won a scholarship of Rs. 2.2 lacs towards my tuition fee at BITS, Pilani.
-
BITS, Pilani - Merit Scholarship (2008-2009)
BITS, Pilani
Merit scholarship for being in the top ten students of the batch during the academic year 2008-2009.
-
PhD admit from the School of Computer Science, Carnegie Mellon University
Carnegie Mellon University
Test Scores
-
Graduate Record Examination (GRE)
Score: 333/340
Quantitative - 169/170
Verbal - 164/170 -
Common Admission Test (CAT)
Score: 99.96 percentile
Admission offers from IIM Bangalore and IIM Calcutta
-
Graduate Management Admission Test (GMAT)
Score: 770/800
Top 1 percentile of test takers across the world.
-
AIEEE - 2008
Score: All India Rank 272
Achieved All India Rank (AIR) 272 out of an estimated 800,000 students.
Approximate percentile = 99.97 -
BITSAT 2008
Score: 374/450
-
GPA - CMU
Score: 3.97/4.0
-
IITJEE - 2008
Score: All India Rank 4743
Out of an estimated 320,000 students.
Languages
-
English
Full professional proficiency
-
Hindi
Native or bilingual proficiency
-
Punjabi
Elementary proficiency
Organizations
-
Triple Nine Society
-
- Present
More activity by Aman
-
It is very important to reduce KV cache memory consumption during long context inference. We introduce ThinK, a method that exploits substantial…
It is very important to reduce KV cache memory consumption during long context inference. We introduce ThinK, a method that exploits substantial…
Liked by Aman Gupta
-
Introducing torchchat 🔥 A lightweight library to run LLMs locally across mobile, desktop and laptops powered by PyTorch. Learn more:…
Introducing torchchat 🔥 A lightweight library to run LLMs locally across mobile, desktop and laptops powered by PyTorch. Learn more:…
Liked by Aman Gupta
-
🍎 Post-training in the Apple Intelligence paper - Two models: - On-device: ~3B param with task-specific LoRA adapters - Server: ~70B (my…
🍎 Post-training in the Apple Intelligence paper - Two models: - On-device: ~3B param with task-specific LoRA adapters - Server: ~70B (my…
Liked by Aman Gupta
-
Efficiently inferencing and fine-tuning massive models like Llama 3.1 405B requires a synthesis of multiple memory optimizations, parallelism and…
Efficiently inferencing and fine-tuning massive models like Llama 3.1 405B requires a synthesis of multiple memory optimizations, parallelism and…
Liked by Aman Gupta
-
🎲 Visualize and interact with LLM decoding strategies LLM sampler is a cool website that allows you to see the effect of temperature, top K…
🎲 Visualize and interact with LLM decoding strategies LLM sampler is a cool website that allows you to see the effect of temperature, top K…
Liked by Aman Gupta
-
Excited to see that the industry at large is recognizing the importance of graph-based RAG. At Glean, this is how we've always approached making…
Excited to see that the industry at large is recognizing the importance of graph-based RAG. At Glean, this is how we've always approached making…
Liked by Aman Gupta
-
Excited to be in Vienna this week to give a talk at the 2nd iteration of the #WANT Workshop on Efficient Training at [ICML] Int'l Conference on…
Excited to be in Vienna this week to give a talk at the 2nd iteration of the #WANT Workshop on Efficient Training at [ICML] Int'l Conference on…
Liked by Aman Gupta
-
PyTorch 2.4 is live 🙌 Featuring: - Python 3.12 support for torch.compile - AOTInductor Freezing for CPU - New Higher-level Python Custom Operator…
PyTorch 2.4 is live 🙌 Featuring: - Python 3.12 support for torch.compile - AOTInductor Freezing for CPU - New Higher-level Python Custom Operator…
Liked by Aman Gupta
-
Ever since the #Budget was presented by our hon'ble Finance Minister, Ms. Nirmala Sitharaman, much of social media has become a long tirade for it or…
Ever since the #Budget was presented by our hon'ble Finance Minister, Ms. Nirmala Sitharaman, much of social media has become a long tirade for it or…
Liked by Aman Gupta
-
Llama 3's paper is full of cool insights. Here's how they filtered out bad instruction samples. - Quality is evaluated by a reward model and an…
Llama 3's paper is full of cool insights. Here's how they filtered out bad instruction samples. - Quality is evaluated by a reward model and an…
Liked by Aman Gupta
-
Why do 16k GPU jobs fail? The Llama3 paper has many cool details -- but notably, has a huge infrastructure section that covers how we parallelize…
Why do 16k GPU jobs fail? The Llama3 paper has many cool details -- but notably, has a huge infrastructure section that covers how we parallelize…
Liked by Aman Gupta
-
Register for our upcoming Campfire Conversation on Friday, July 26th at 12 noon PT. Join us for this open and informal event as we dive into the…
Register for our upcoming Campfire Conversation on Friday, July 26th at 12 noon PT. Join us for this open and informal event as we dive into the…
Liked by Aman Gupta
-
Exciting News, Everyone! We are gearing up to recruit for our Summer 2025 Interns soon! To ensure your profile stands out during our outreach…
Exciting News, Everyone! We are gearing up to recruit for our Summer 2025 Interns soon! To ensure your profile stands out during our outreach…
Liked by Aman Gupta
-
At #ICML2024? Come drop by the Netflix booth in hall B and learn more about what we're doing and and our open positions. We are hiring for many types…
At #ICML2024? Come drop by the Netflix booth in hall B and learn more about what we're doing and and our open positions. We are hiring for many types…
Liked by Aman Gupta
Other similar profiles
Explore collaborative articles
We’re unlocking community knowledge in a new way. Experts add insights directly into each article, started with the help of AI.
Explore MoreOthers named Aman Gupta in United States
-
Aman Gupta
Talent Acquisition Specialist
-
Aman Gupta
Vice President of Acquisitions at Starwood Capital Group
-
Aman Gupta
-
Aman Gupta
151 others named Aman Gupta in United States are on LinkedIn
See others named Aman Gupta