Tatev Aslanyan

Senior Machine Learning Researcher

With a Master's degree in Econometrics, I bring 9 + years of robust international hands-on technical experience in Machine Learning, Data Science, and AI from Europe, US and Canada. High-impact global projects across Ecommerce, Recommender Systems, Search Engines, Marketing and more, with a focus on using Python, SQL, PySpark and cloud architectures like AWS. My skills extend to building and training Machine Learning and Deep Learning models in Python, complex SQL querying, cloud architecture, and econometric modelling. I am recognized for my analytical capabilities, statistical and mathematical background, and first-authored publications in top-tier US journals centered on NLP &amp; ML. An independent thinker. I thrive both autonomously and as a team player, complemented by my strong communication and collaboration abilities. Worked in Tech, as Data Scientist in ML and AI in both product-based and consultancy settings with a strong business sense, and collaborating with Tech Leads, UX designers, Senior Management, Product Managers, and Business Directors.🔍 Operate:Specialties: Machine Learning, Artificial Intelligence, NLP, Deep Learning, Advanced AnalyticsHave also Experience: Data Engineering, Product Data Science, LLMs(Bert)Innovation: Developed a recognized Data Science Bootcamp; Published first-authored papers in AI and NLP research at ACM🔗 Engage &amp; Explore:📘 Personal Portfolio: tatevaslanyan.com🚀 Data Science Bootcamp: lunartech.ai 💻 YouTube Channel: https://youtube.com/@LunarTech_ai?si=Xka8dNPM8Y2cB3S9📧 "The Data Science and AI Newsletter": https://tatevaslanyan.substack.com/💻 Github Code and Projects: https://github.com/TatevKaren✍️ Blog: medium.com/@tatevkarenaslanyan📆 Schedule a Call: tatevaslanyan.com/contact/Explore numerous Free Resources and Technical Code/Materials on my LinkedIn, Personal Website and GithubFeel free to explore the links above or drop me a message directly here on LinkedIn

Co-Founder

Co-Founder of LunarTech, and creator of LunarTechs "The Ultimate Data Science Bootcamp" ranked as one of the Best Data Science Bootcamps of 2023 [ITPro] featured on Forbes, TechTimes, and other international journals.

I created the following courses as part of LunarTech initiative:
🔘 Data Science Career Guide Course
🔘 Fundamentals to Statistics Course
🔘 Fundamentals to Machine Learning Course
🔘 A/B Testing Course
🔘 Introduction to NLP Course
🔘 Python for Data Science Course
🔘 Applied Data Science Course

No matter how many years of experience you have, if you don’t know how to treat people, it’s not “leadership.” It’s just a résumé. 🤝

- Skill gets you in the room.
- Character decides whether people stay in the room with you.

The older I get, the clearer it becomes: most “high performers” don’t fail because they aren’t smart. They fail because they create fear, ego, and silence around them.

If you want a simple leadership checklist:
- Speak with respect (even when you’re stressed)
- Give credit loudly, give feedback privately
- Keep your word
- Make people feel safe to tell you the truth and be free

Your title is temporary. Your reputation is permanent.

Don’t forget to follow LunarTech on social media for more AI & data science. 🤝

And if you want to switch careers (or level up fast): #LunarTechAcademy helps you become job-ready in AI + Data Science—with formal certifications in Data Science, AI Engineering, and AI for Executives.

Founder and CEO @ LunarTech | AI Engineer and Data Scientist | Seen on Forbes, Yahoo, Entrepreneur | Empowering Enterprises with Data Science and AI

There’s a popular idea floating around that early-stage founders should “activate their networks” to hire—former colleagues, friends of friends, people vouched for by angels or VCs. The argument is that using your personal network is faster, more efficient, and shows strength in “relationship capital.” I disagree—completely.

Early in our journey, I hired someone based on this exact mindset. A shared connection, someone from a familiar circle, vouched for their values and “good character.” That person turned out to be one of our worst hires.

No matter how well-meaning the referral was, the reality is:
▶️ Trust-based hiring leads to false confidence.
▶️ Familiarity becomes a substitute for skill.
▶️ And the cost of fixing that mistake is far higher—in time, in morale, and in culture—than simply hiring right through a proper, merit-based process.

I’ve seen this pattern repeat itself in startups, corporations, FAANG pipelines, and beyond. It breeds a dangerous environment:
• Some feel untouchable because they “know someone.”
• Others feel they weren’t hired on equal footing.
• Office politics starts to take root.
• And great talent—people who could genuinely contribute—is left on the shelf.

Culture gets damaged before it’s even built.

But it doesn’t stop at startups. This kind of network-based favoritism exists everywhere: in medicine, in justice systems, in public offices, and in every field where trust is misplaced and merit is ignored. You end up with unqualified people making decisions they have no right to make—while truly skilled, well-mannered, and capable individuals are unemployed or underused.

This isn’t just a hiring inefficiency—it’s a systemic failure. And we see the consequences daily.

So no, it’s not “faster” to hire from your network. It’s costlier. And it contributes to a long-term culture of incompetence.

If you’re a founder—even in the early days—build your brand. Let your company attract talent. Evaluate fairly. Hire through process, not connections. Trust your own judgment—not your friends’ opinions.

Merit > networks.
Process > shortcuts.
Earned roles > granted roles.

That’s the culture we’ve built—and it works.

This is my opinion, shaped by experience. And I stand by it! 

#LunarTech #culture #startups

No alternative text description for this image

We’re #hiring a Software Engineering Apprentice (AI Engineering Team) — Armenia 🇦🇲

This 6-month apprenticeship is for recent graduates and career switchers ready to build real production systems and AI-related projects.

Key details
- Remote-first for now (Armenia)
- 25 hours/week | Unpaid internship
- English B2 minimum (C1 preferred)
- Opportunity to be considered for a full-time offer after the program (not guaranteed)

What you’ll do
Build web + mobile apps, implement UI from UX designs, integrate - APIs/services, and help productionize AI modelsalongside AI engineers, UX designers, bioinformatics professionals, and data scientists across Western Europe, North America, Latin America, and Eastern Europe.

Tech (examples): Next.js/TypeScript, React Native or Flutter, FastAPI/NestJS, PostgreSQL/Supabase (pgvector), Docker, CI/CD.

👉 Apply via LinkedIn or send CV + GitHub/portfolio + short note to me! Comment interested below if the link is filled!

#aiengineering #hiring #lunartech

Have you ever been in this situation?

You’re sitting with a PDF or a handbook — an e-book for learning, a research paper, or a medical / legal document — and it’s in a language you don’t speak.

You need it in another language. Fast. And accurately.

You try Google and ChatGPT Translate.
You try copy–paste.
Everything breaks.

Formulas disappear.
Tables fall apart.
Images shift.
Footnotes get lost.
And the structure that actually matters is gone.

This is exactly why we built Babel.

Babel is a layout-faithful handbook + PDF translation engine.
It translates complex documents into your desired language while keeping everything else the same.

That means:
tables stay tables,
formulas (even complex math) stay formulas,
images remain in place,
headers, footers, references, and layout are preserved.

Only the language changes.

Yesterday, we open-sourced Babel and published our full test results.

We translated highly complex research papers (90–110 pages) across:
Arabic, Chinese, Spanish, French, Italian, Japanese, Dutch, Portuguese, Russian — and more.

What used to take weeks and thousands of dollars
now takes a few minutes.

We keep seeing the same use cases:
Students translating textbooks and handbooks for learning.
Researchers translating papers between Chinese ↔ English (and beyond).
Healthcare teams translating protocols and technical documentation.
Universities localizing academic material.
Anyone with an e-book who simply wants it in their own language.

If you have a research paper in Chinese and need it in English — Babel does it.
If you have an English handbook and want it in Armenian — Babel does it.

(For scanned documents, Babel Extreme handles translation + reconstruction.)

We open-sourced Babel because we believe language should never be a barrier to knowledge.

The tech behind it combines document-structure parsing, layout reconstruction, and AI-powered translation — built specifically for PDFs and handbooks, not generic text.

You’re free to:
use it, contribute, share it, and build with us.

Open source + tests on LunarTech Github

If you’ve ever struggled with PDFs across languages — this one’s for you.

About year ago I joined association. In my first presentation I mentioned my education and work experience outside🇦🇲 

A man objected. Then he asked something that’s actually worth answering:

“Why do we always hear about Armenian founders succeeding outside Armenia?”

So here’s my answer—based on watching, listening,quietly observing how “tech” works here

#Why Armenians succeed abroad

Because outside Armenia, there is usually higher chance that real value gets rewarded: competence, delivery, ethics, consistency, dignity. Not perfectly, not everywhere—but more often than in the current systems in Armenia

Inside Armenia, success can be filtered through a different logic: permission, not merit

let’s be clear: this is not about being extroverted. It’s not “go to more events.” It’s not “network harder.” What I’m describing is when ecosystem behaves like closed trust circle—where access depends on belonging, protection, loyalty chains more than on what you build. People are even afraid to say something because of the fear culture 

You see it in patterns people whisper about:
	•	same organizations & faces “rotating” through grants, sponsorships, visibility
	•	“community” platforms acting like gatekeepers, not enablers
	•	strong sexism & double standards still shaping who gets the spotlight and who gets “managed”

And if you stand up for yourself as a women and set boundaries- response isn’t “let’s fix it.”
It’s: “Let’s make YOU the problem.”
—while the “other founders” being showcased are already inside protective bubble. 

#The easiest example: AI-washing

Armenia loves to market “AI.” But you can literally feel the gap between branding and building when you meet the “AI” up close.

You’ll get a call from a so-called Armenian AI assistant, sounding futuristic… then you reply with your first sentence in English—
and suddenly the “AI” becomes a confused script that can’t handle reality.

That’s what happens when a system rewards labels, cheating, incompetence over competence.

#Davos - the uncomfortable reality check 🇦🇲

At Davos, Azerbaijan’s president Ilham #Aliyev said international relations are moving from the “rule of law” to “rule of strength”

If the world is becoming more brutal & competitive which is especially dangerous for Armenia, then in the very near future, fake companies, fake products, fake people, fake reality won’t save anyone

- Only real products, real progress, & real founders will

And that’s why this matters. The game being played right now in Armenia —where branding replaces building, - loyalty replaces merit—is a dangerous one in 2026 (some at the top who are involved know this - that’s why are perping their foreigner passports & exit plan) 

I’m not posting this to shame Armenia. I’m posting it because Armenia has real talent—and real builders including us deserve real ecosystem where value is the admission ticket, not “being trusted” by a cult.

LETS make real products, real connections & real impact! 🇪🇺🇦🇲

People love to blame “Gen Z” for poor work ethic or professionalism. I don’t buy it.

At LunarTech, we’ve worked with some outstanding Gen Z collaborators — curious, fast, reliable, and genuinely impactful. We’ve also seen unprofessional behavior from people across different ages, genders, and nationalities. In other words: it’s not generational. It’s individual.

A recent example: we opened a role for a grant writer. One applicant used most of the conversation to push “branding” instead (not what the role was for). Then came the pitch: “I’ll fix a couple of typos on your homepage for free… and you give me a testimonial.”

That’s not value creation. That’s a shortcut to borrow credibility — and to take credit for a brand founders build through years of daily work, consistency, and delivery.

The takeaway for founders and candidates:
	•	Respect the scope of the role
	•	Don’t waste time with misaligned pitches
	•	Don’t try to claim outsized credit for small cosmetic edits
	•	Be consistent: intentions show up in follow-through

Work ethic comes down to the person — how they were raised, their professional standards, and their intentions. Not their generation.

#Hiring #Startups #Professionalism #Integrity #WorkEthic #BrandBuilding

We’re #hiring: Grant Writer / Grant Strategist (Remote, Part-time)

LunarTech is looking for an experienced grant strategist to lead high-value funding efforts across AI / Deep Tech, with a focus on healthcare, education, and applied innovation.

- 100% commission-based (success-fee only) — no retainer
- Minimum 10 hours/week
- EU / US / Middle East funding experience is a plus
- Comfortable owning the full lifecycle: opportunity → strategy → compliant submission

If interested, DM me and I’ll share the role + survey.

Grant Writer / Grant Strategist (Remote, Part-Time, Commission-Based)

I’m a food industry specialist and supply chain / manufacturing strategist who helps leaders connect long-term industry shifts with near-term operational wins. Over the past 20+ years, I’ve worked across strategy, operations, and transformation with one consistent focus: turning complexity into measurable results. Today, I’m a Partner at Value Gene, supporting food manufacturers and distributors with programs that deliver tangible ROI while building the capabilities needed for what’s next.

My work sits at the intersection of food, manufacturing, and emerging technology—especially how AI and humanoid robotics, along with demographic, climate, and health shifts, will reshape how companies operate and compete.

Ozan Ozaskinli

Over five years, U.S. food and beverage growth followed a pattern that most companies still underestimate.

In the last decade, the majority of category growth came from two ends of the market: Small Brands and Store Brands. Growth concentrated rather than spreading evenly across categories.

Small Brands captured share in categories where consumers actively evaluate benefits. Just in the last 5 years, in health- and function-led segments, they gained five to six points of share, while large national brands lost ground. These brands are explicit about who they serve and the specific benefit they deliver. They focus on defined jobs such as functionality, protein, or lifestyle support. When consumers pause and compare options, brands with a clear purpose and a specific benefit gain share.

Store Brands grew for a different reason. In everyday categories with limited perceived differentiation, consumers prioritized value, reliability, and availability. Store Brand share increased from about 14.5% to roughly 16% over the same period. In regular products, they account for over 25% of volume. Inflation widened price gaps and accelerated Store Brand adoption, with behavior persisting even as pricing pressure eased.

Brands losing ground tend to sit in the middle. They price above Store Brands while offering weaker differentiation than focused challengers. Many rely on incremental renovation, assuming scale and distribution provide protection. The data shows continued share erosion in this middle position, even inside categories that expand.

Growth today rewards clarity of intent. Products that commit to a clearly defined value proposition gain momentum. Products that attempt to combine value positioning and premium differentiation struggle to progress.

Food & Consumer Goods Advisor | Consulting Partner | Ex-BCG

As I look back on 2025 in food industry, what stands outmost was how many changes people still described as temporary. 

I kept hearing the expectation that markets would stabilize and patterns would revert. My experience this year pointed in a different direction.

Fragility surfaced across the food system quickly this year. Small shocks created outsized effects. Trade shifts moved volumes across borders within months. Input volatility showed up faster than planning cycles could absorb. Teams with tightly optimized systems felt the strain first.

We also saw health move from aspiration to baseline. Consumers treat it as an expectation. Products that simply improved nutrition struggled for attention. Growth showed up when companies addressed specific needs such as functionality, or longevity. This reinforced something I wrote earlier this year: incremental change feels safe internally, yet it rarely moves demand.

I was also struck by how decision speed shaped outcomes. Teams that accepted uncertainty and made directional choices adapted faster. This showed up across portfolio decisions, sourcing shifts, and operating model changes. Waiting for perfect information has a cost.

In 2025, the food industry entered a period where assumptions expire quickly. Leaders who acknowledge that reality early gain room to maneuver. Those who expect a return to familiar patterns face a narrower set of options.

I’m genuinely proud to share what we just published: How Humanoids Will Reshape Food Manufacturing.

Humanoids won't be widely used in factories for a few more years. Still, I believe it's important to start thinking about them in a clear and detailed way now.  Companies that prepare early will have an advantage. That’s exactly why we created this.

We took a step back to answer some key questions: Where does the technology stand today? What changes are coming in the next few years? And what will 'humanoid transformation' actually look like in a food plant?

The result is a 70+ page deep dive. If you prefer audio, we also turned it into a podcast: Value Gene Insight Conversations (available on YouTube, Spotify, Apple Podcasts, and other major platforms).

📄Read the full article: https://lnkd.in/gGzhx5nJ 🎧Listen to the podcast: https://lnkd.in/gMu7jfQT

This topic has many unknowns, so I’d really appreciate your insights and feedback. What resonates, what doesn’t, and what is underestimated in this discussion? Your feedback will shape future analysis.

Humanoids Food Manufacturing

When we started working on How Humanoids Will Reshape Food Manufacturing, I had certain assumptions. But as we dug into the analysis, my perspective changed. Three things in particular surprised me.

1) The timeline may move much faster than most people think. The current trajectory still suggests “a decade away.” But there is credible evidence that it could take closer to 6 years to reach a point where humanoids can handle a large share of the tasks currently performed by people in plants. More investment and advances in AI could speed things up even more.

2) The payback is real, even early on. Even if early-generation humanoids cost more and aren’t as efficient, the numbers can still work. In our scenarios, payback ranges from 18 to 22 months, and could drop to 7 to 10 months as costs go down and performance improves. This could quickly become a real investment for many companies.

3) Change management will be a different game. This might be the hardest part. Many transformation programs work because they make jobs easier, safer, and more valuable. Humanoids are different. They will create fewer jobs than they replace, and pretending otherwise will backfire. My takeaway is that companies need to be clear and, honest, and strong about change from the start, with transparent plans for workforce transition. Otherwise, the disruption inside the company could be even bigger than the technology shift itself.

If you haven’t checked the piece yet, here are the links: 📄Article: https://lnkd.in/gGzhx5nJ 🎧Podcast (Value Gene Insight Conversations): https://lnkd.in/gMu7jfQT

After reading or listening, I’d really appreciate your feedback. What do you agree with? What do you disagree with? And what do you think we missed?

taplio

The Battle

Tatev Aslanyan

Ozan Ozaskinli

Followers

Posts

Engagements

Likes

Performance ScoreTatev Aslanyan's LinkedIn Performance

Performance ScoreOzan Ozaskinli's LinkedIn Performance

Ready to grow your LinkedIn presence?

taplio

The Battle

Tatev Aslanyan

Ozan Ozaskinli

Followers

Posts

Engagements

Likes

Performance Score.css-goub6s{font-size:var(--chakra-fontSizes-md);font-weight:500;color:var(--chakra-colors-gray-600);display:block;margin-top:var(--chakra-space-1);}Tatev Aslanyan's LinkedIn Performance

Performance ScoreOzan Ozaskinli's LinkedIn Performance

Ready to grow your LinkedIn presence?

Performance ScoreTatev Aslanyan's LinkedIn Performance