Sign In

Hrag Balian

Engineering | ML | Analytics | Data | Ops | Cloud

Hrag Balian is a seasoned professional with over 12 years of experience in technical roles focused on building Data and Machine Learning (ML) products.

He specializes in operationalizing business value into efficient and cost-effective cloud systems, collaborating cross-functionally from ideation to product launch, and leading high-performing teams to drive successful outcomes.

His expertise spans various areas including Cloud technologies such as AWS and GCP, programming languages like Python, project management tools like Confluence, Jira, and Figma, and statistical and ML frameworks such as PyTorch, Scikit-learn, GLMs, LLMs, and XGBoost.

Hrag is well-versed in AI and Vector Databases such as Llamaindex, Langchain, and LanceDB, as well as API & Web frameworks like Reflex, FastAPI, Streamlit, Dash, and Ops/Orchestration tools like Lambda, ECS, Airflow, Terraform, Serverless Framework, Github Actions, Docker, and boto3.

His experience includes working with Rectangular Data using tools like SQLAlchemy, PostgreSQL, Athena, MongoDB, and Spark, showcasing his diverse technical skill set.

Hrag Balian's educational background is extensive, having studied Ph.D. in Sociology at Columbia University, Post-Doctoral Associate in Organizational Behavior at Yale School of Management, MA in Sociology/Social Statistics at McGill University, and BA in Sociology at McGill University.

Throughout his career, Hrag has held key positions in reputable organizations, serving as the Former Sr. Director of Analytics and Data Engineering at Material, Former Director of Data Science at LRW, Director of Marketing Science at Lieberman Research Worldwide, Senior Manager at Lieberman Research Worldwide, Manager of Marketing Science at Lieberman Research Worldwide, Post-Doctoral Associate at Yale School of Management at Yale University, and Doctoral Student in Sociology + Post-Doctoral Fellow at the Paul Lazarsfeld Center at Columbia University.

Highlights

Nov 10 · linkedin

There is a man-made humanitarian crisis unfolding in Gaza right now — over 10,000 people have been killed, with countless more injured. Lack of clean water and access to food puts hundreds of thousands more at risk every second.

Call your elected leaders to call on a ceasefire. Tell them that you stand for human rights and want to see an immediate end to the bombardment of Palestinians in Gaza. It’s the right thing to do.

If you’re in California:

Senate:  Alex Padilla: (202) 224-3553 Laphonza Butler: (202) 224-3841

Find your Congress representative here: https://lnkd.in/g29PyEnP

If you’re still wondering if ceasefire is the right thing to do, consider the following two facts:

  • The 121 member states of the United Nations who voted for the “humanitarian truce” resolution ES-10/21 represent 60% of the global population. These countries represent only 43% of the global GDP.

  • There is unity among hundreds of global and regional human rights organizations in calling on Israel for a cessation of hostilities, from well-known ones like Human Rights Watch, Amnesty International, and the World Food Program to lesser known ones like CIVIC and Global Voices.

What conclusion can we draw from these facts ?

  1. The global majority supports ceasefire. 
  2. The less powerful support ceasefire. 
  3. Those who know and understand violence support ceasefire.
  4. The people and organizations who can reasonably claim a moral high ground support ceasefire.

From a humanitarian standpoint, ceasefire is the right thing to do. Gaza needs a ceasefire to receive aid at a scale that it hasn’t yet seen, and that it desperately needs.

#ceasefirenow

Jul 12 · linkedin

I’ve been experimenting with tooling within the emerging AI ecosystem, like LangChain, Chroma, Pinecone, and others, to orient myself to the evolving AI landscape. In working with these tools, and seeing how to leverage them in differentiated ways, it has hardened my stance that a successful AI strategy would need to be coupled with a comprehensive Data strategy to be fully adaptive. An AI strategy without a Data strategy, or vice versa, could be fine in the short term but not really have much long term benefits to an organization.

Personally, I need simple frameworks to organize my thoughts. This 2x2 represents my initial thinking on this.

Let’s call organizations that do neither (lower left quadrant) “non-adaptive”, as new technologies are ignored by the organization, leaving all legacy systems and processes intact, with little to no meaningful resources put toward improving either product or process.

An “opportunistic” organization is one that develops an AI strategy without a Data strategy (upper left quadrant). The focus here will be less on product and more on process, where the organization leans heavily on baseline models that every other organization has access to to deliver the same things they currently do, but faster or more efficiently.

A “myopic” organization is one that has invested in a Data strategy without an AI Strategy (lower right quadrant). These organizations will focus on their existing products, and existing roadmaps, without putting in the resources to think through how AI can help them build, or how it can change the nature of what they build.

An “adaptive” organization is one that puts AI and Data on an equal footing (upper right quadrant). These organizations will lean in to their proprietary data to differentiate their product on their data, and to fuse AI with every aspect and feature of their product by leveraging their unique data, thereby creating a moat.

There are shades of grey between these organizations, and organizations can move through this space, but this framework helps me at least think through and highlight what I believe would be the outcome of investing in one or the other, or both, or neither.

I’m anticipating seeing many older organizations be “non-adaptive”, with most of the remaining larger organizations as being “opportunistic” or “myopic”.  The few fully “adaptive” organizations that offer products that fully integrate an AI strategy with a Data strategy will be best positioned to differentiate and outperform others.

Needless to say, these are the early days of AI, but decisions (and indecisions) made now will likely have lasting consequences.

#generativeai #ai #strategy #data #datastrategy #aistrategy

Hrag Balian
Add to my network

Location

Los Angeles Metropolitan Area