Hey, I'm Saksham.
I'm an AI & analytics consultant at EXL. Day to day that means data pipelines, models and LLM-powered tools, plus the less glamorous job of explaining to non-technical people what they actually do and whether they're worth it.
I spend a lot of time on AI Twitter, so every week I turn the noise into Keeping Up With AI: new models, tools that shake up a sector, real breakthroughs, and the occasional bit of tea on the people building all of it. Written for professionals who want to keep up without a CS degree.
Skills & Stack
GenAI & LLMs
Claude API
Gemini
LangChainML & Deep Learning
TensorFlow
Keras
scikit-learnData & Analytics
Python
PySpark
Pandas
NumPyCloud & Engineering
Firebase
Git
JavaScript
React Native
CursorExperience
2026 - Present
Consultant, AI & Analytics
EXL
Working on credit strategy for one of America's largest banks: SQL and PySpark data work, predictive models, and LLM applications such as natural-language-to-SQL and retrieval over internal knowledge. Part of the job is translating between the technical team and the people who sign off on it.
Feb. '26 - Present
Valen: Recovery Intelligence Fitness App
Independent Project
Built and shipped a React Native fitness app, now live in production, with a fatigue-based recovery engine. Designed it end to end, learned release pipelines and testing along the way, and used its own data as a sandbox for ML experiments: validating the recovery formula against how sessions actually felt, and training a classifier to tag exercises by muscle zone. Featured in Android Planet's Best Apps of the Week.
Jan. '25 - Jun. '25
Worked on the LLM layer of an AI ad-tech product: prompt structure, context pruning and API efficiency on OpenAI models, cutting latency and token usage by roughly 12%. Also helped ship a trimmed-down production version after the CTO left.
Aug. '24 - Oct. '24
Research Intern (Remote)
University of New South Wales
A remote research stint on how consumers in a new market choose between products. It taught me how to run a study, read messy survey data and turn it into a recommendation, which is where my comfort with data started.
Aug. '23 - Dec. '23
Founder's Office Intern
Edviron
A first look at how an early-stage startup actually runs: operations, onboarding and the unglamorous work behind a small team. It showed me how much of a company is process, and how much of that process software can take over.
Sep. '20 - Jul. '21
Freelance Technical Writer
Self-employed
Started writing about consumer tech at 16, researching and explaining gadgets to general readers. Same skill I use now: take something technical and make it clear to someone who has better things to do.