Aurimas Griciūnas

Senior Solutions Architect

Follow me to learn about Data Engineering and ML Systems | Author of SwirlAI Newsletter | Public Speaker

Tired of all the noise surrounding AI space in recent years? Having a hard time keeping up with all of the news in the field and distinguishing what is really important? Follow me and subscribe to my newsletter to get actionable insights on building production ready AI Systems and keep track of only the important signal of what is happening in the AI world.

With over a decade of experience in Data Science, AI/ML Engineering, and MLOps, I’m also dedicated to helping practitioners level up their skills in MLOps, Data Engineering, Machine Learning, and related fields. I share my knowledge in a clear and accessible way to make learning straightforward and actionable.

If you want to partner with SwirlAI and get your product in front of 40K+ AI/ML/Data Engineers and Leaders via our Newsletter or 135K+ AI/ML/Data Engineers and Leaders via LinkedIn, contact me directly.

Head of Product

Leading Product Strategy.

Fusion of 𝗥𝗔𝗚 (Retrieval Augmented Generation) and 𝗖𝗔𝗚 (Cache Augmented Generation). How can you benefit from it as AI Engineer?

Let’s see what it looks like and what additional considerations should be taken into account.

Here are example steps to implement CAG + RAG architecture:

𝘋𝘢𝘵𝘢 𝘗𝘳𝘦𝘱𝘳𝘰𝘤𝘦𝘴𝘴𝘪𝘯𝘨:

𝟭. We use only rarely changing data sources for Cache Augmented Generation. On top of the requirement of data changing rarely we should also think about which of the sources are often hit by relevant queries. Once we have this information, only then we pre-compute all of this selected data into a KV Cache of the LLM. Cache it in memory. This only needs to be done once, the following steps can be run multiple times without recomputing the initial cache.
𝟮. For RAG, if necessary, precompute and store vector embeddings in a compatible database to be searched later in step 4. Sometimes simpler data types are enough for RAG, a regular database might suffice.

𝘘𝘶𝘦𝘳𝘺 𝘗𝘢𝘵𝘩:

We can now utilise the preprocessed data.

𝟯. Compose a prompt including user query and the system prompt with instructions on how cached context and retrieved external context should be used by the LLM.
𝟰. Embed a user query to be used for semantic search via vector DBs and query the context store to retrieve relevant data. If semantic search is not required, query other sources, like real time databases or web.
𝟱. Enrich the final prompt with external context retrieved in step 4.
𝟲. Return the final answer to the user.

𝘚𝘰𝘮𝘦 𝘊𝘰𝘯𝘴𝘪𝘥𝘦𝘳𝘢𝘵𝘪𝘰𝘯𝘴:

➡️ Context window is not infinite and even while some models boast enormous context window sizes, the needle in the haystack problem has not yet been solved so use available context wisely and cache only the data you really need.

✅ For some business cases, specific datasets are extremely valuable to be passed to the model as cache. Think about an assistant that has to always comply with a lengthy set of internal rules stored in multiple documents. 
✅ While CAG has been popularised for Open Source just recently, it is already viable for some time via Prompt Caching features in OpenAI and Anthropic APIs. It is really easy to start prototyping there.
✅ You should always separate hot and cold data sources, only use cold (data that changes rarely) in your cache, otherwise the data will go stale and the application will go out of sync.

❌ Be very careful about what you cache as the data will be available for all users to query.
❌ It is very hard to ensure RBAC for cached data unless you have a separate model with its own cache per role.

Learn how to apply the technique in my End-to-end AI Engineering Bootcamp. Next cohort kicking off on January 12th: https://lnkd.in/dagWE5r3

🎁 Code LASTCHANCE for 10% off.

Founder @ SwirlAI • Ex-CPO @ neptune.ai (Acquired by OpenAI) • UpSkilling the Next Generation of AI Talent • Author of SwirlAI Newsletter • Public Speaker

𝗠𝗖𝗣 plus 𝗔𝟮𝗔, here is how they complement each other 👇

Protocol wars continue to rage, let's understand how Googles A2A (Agent2Agent) protocol is different from MCP and how they complement each other (read till the end).

𝘔𝘰𝘷𝘪𝘯𝘨 𝘱𝘪𝘦𝘤𝘦𝘴 𝘪𝘯 𝘔𝘊𝘗:

𝟭. MCP Host - Programs using LLMs at the core that want to access data through MCP.

❗️ When combined with A2A, an Agent becomes MCP Host.

𝟮. MCP Client - Clients that maintain 1:1 connections with servers.
𝟯. MCP Server - Lightweight programs that each expose specific capabilities through the standardised Model Context Protocol.
𝟰. Local Data Sources - Your computer’s files, databases, and services that MCP servers can securely access.
𝟱. Remote Data Sources - External systems available over the internet (e.g., through APIs) that MCP servers can connect to.

𝘌𝘯𝘵𝘦𝘳 𝘈2𝘈:

Where MCP falls short, A2A tries to help. In multi-Agent applications where state is not necessarily shared

𝟲. Agents (MCP Hosts) would implement and communicate via A2A protocol, that enables:

➡️ Secure Collaboration.
➡️ Task and State Management.
➡️ User Experience Negotiation.
➡️ Capability discovery - similar to MCP tools.

𝗛𝗼𝗻𝗲𝘀𝘁 𝘁𝗵𝗼𝘂𝗴𝗵𝘁𝘀:

❗️ Open protocols for Agent communication are important but they are just a piece of the picture, there will be a need for standards that govern all of the existing protocols and other missing pieces. E.g.

❓ How do we standardise tracing and Observability in multi-agent IOA systems?
❓ How do we retain the identity of the running job of an AI Agent instance if the communication standard is not unified in different parts of the pipeline?
❓ ... 

✅ More on this in my future posts, stay tuned!

Let me know your thoughts in the comments. 👇

A breakdown of 𝗗𝗮𝘁𝗮 𝗣𝗶𝗽𝗲𝗹𝗶𝗻𝗲𝘀 𝗶𝗻 𝗠𝗮𝗰𝗵𝗶𝗻𝗲 𝗟𝗲𝗮𝗿𝗻𝗶𝗻𝗴 𝗦𝘆𝘀𝘁𝗲𝗺𝘀 👇 And yes, it can also be used for LLM based systems!

It is critical to ensure Data Quality and Integrity upstream of ML Training and Inference Pipelines, trying to do that in the downstream systems will cause unavoidable failure when working at scale.

There is a ton of work to be done on the Data Lake or LakeHouse layer. 𝗦𝗲𝗲 𝘁𝗵𝗲 𝗲𝘅𝗮𝗺𝗽𝗹𝗲 𝗮𝗿𝗰𝗵𝗶𝘁𝗲𝗰𝘁𝘂𝗿𝗲 𝗯𝗲𝗹𝗼𝘄.

𝘌𝘹𝘢𝘮𝘱𝘭𝘦 𝘢𝘳𝘤𝘩𝘪𝘵𝘦𝘤𝘵𝘶𝘳𝘦 𝘧𝘰𝘳 𝘢 𝘱𝘳𝘰𝘥𝘶𝘤𝘵𝘪𝘰𝘯 𝘨𝘳𝘢𝘥𝘦 𝘦𝘯𝘥-𝘵𝘰-𝘦𝘯𝘥 𝘥𝘢𝘵𝘢 𝘧𝘭𝘰𝘸:

𝟭: Schema changes are implemented in version control, once approved - they are pushed to the Applications generating the Data, Databases holding the Data and a central Data Contract Registry.

Applications push generated Data to Kafka Topics:

𝟮: Events emitted directly by the Application Services.

👉 This also includes IoT Fleets and Website Activity Tracking.

𝟮.𝟭: Raw Data Topics for CDC streams.

𝟯: A Flink Application(s) consumes Data from Raw Data streams and validates it against schemas in the Contract Registry.
𝟰: Data that does not meet the contract is pushed to Dead Letter Topic.
𝟱: Data that meets the contract is pushed to Validated Data Topic.
𝟲: Data from the Validated Data Topic is pushed to object storage for additional Validation.
𝟳: On a schedule Data in the Object Storage is validated against additional SLAs in Data Contracts and is pushed to the Data Warehouse to be Transformed and Modeled for Analytical purposes.
𝟴: Modeled and Curated data is pushed to the Feature Store System for further Feature Engineering.
𝟴.𝟭: Real Time Features are ingested into the Feature Store directly from Validated Data Topic (5).

👉 Ensuring Data Quality here is complicated since checks against SLAs is hard to perform.

𝟵: High Quality Data is used in Machine Learning Training Pipelines.
𝟭𝟬: The same Data is used for Feature Serving in Inference.

Note: ML Systems are plagued by other Data related issues like Data and Concept Drifts. These are silent failures and while they can be monitored, we can’t include it in the Data Contract.

Let me know your thoughts! 👇

The main 𝗔𝗜 𝗔𝗴𝗲𝗻𝘁 𝗗𝗲𝗽𝗹𝗼𝘆𝗺𝗲𝗻𝘁 𝗧𝘆𝗽𝗲𝘀 👇

It’s no secret that an AI Agent is a piece of software wrapping an LLM or multiple LLMs as a reasoning engine to guide its execution flow. The deployment types for these agents also closely resemble those of software or ML applications, here are the main ones:

➡️ 𝗕𝗮𝘁𝗰𝗵:

- You run your Agents as a part of ETL/ELT Process on a given schedule.
- An agent takes the form of a CLI script most of the time and is scheduled to run periodically via tools like Apache Airflow.
- You consume the accumulated work, connect to external context and tools if needed and save the agent's results to a batch storage.
- Computation results can be loaded into a real-time storage and used for real time applications as well.

➡️ 𝗘𝗺𝗯𝗲𝗱𝗱𝗲𝗱 𝗶𝗻 𝗮 𝗦𝘁𝗿𝗲𝗮𝗺 𝗔𝗽𝗽𝗹𝗶𝗰𝗮𝘁𝗶𝗼𝗻:

- You apply your agentic logic as a part of a Stream Processing Pipeline.
- While work events are continuously piped through your Streaming Data Pipelines, an agent application applies computation on this data and returns it to the system - most likely another Streaming Storage.
- This deployment type can also easily connect to any required external context storage or tools.
- Outputs can be consumed by multiple applications subscribing to the Inference Stream.

✅ This is an ideal deployment type for Ambient Agents, they can listen for events flowing through the system and pick up the ones dedicated to them.

➡️ 𝗥𝗲𝗮𝗹 𝗧𝗶𝗺𝗲:

- You expose your agent as a Backend Service (REST or gRPC).
- This agent retrieves any required external context from external context stores as well as utilising tools available to it.
- Computation can be requested by any application in real time.

➡️ 𝗘𝗱𝗴𝗲:

- You embed your agent's logic directly into the application that runs on a user device.
- This method provides the lowest latency and improves privacy.
- Data in most cases is generated and lives inside of the device significantly improving the security.

Learn how to apply these in practice in my End-to-end AI Engineering Bootcamp: https://lnkd.in/dagWE5r3

Kicking off next Monday, use code LASTCHANCE for 10% discount!

What types of deployments are you mostly working on? Let me know in the comments! 👇

Founder @ SwirlAI • UpSkilling the Next Generation of AI Talent • Author of SwirlAI Newsletter • Public Speaker

𝗔𝗜 𝗢𝗯𝘀𝗲𝗿𝘃𝗮𝗯𝗶𝗹𝗶𝘁𝘆 is a must have in your tool belt as an AI Engineer. 𝗧𝗿𝗮𝗰𝗶𝗻𝗴 sits at the core of it, why is it important?

Tracing and instrumentation of software have been around for decades now. With AI systems resembling regular software even more, we are now moving the practice here as well (with a few key differences).

Let’s look into the process of tracing from a perspective of a naive RAG system.

𝘍𝘦𝘸 𝘥𝘦𝘧𝘪𝘯𝘪𝘵𝘪𝘰𝘯𝘴:

𝘼) An Orchestrator in the GenAI system application is the central piece of software that orchestrates the end-to-end process. Think of apps using LangChain, LlamaIndex or Haystack.
𝘽) Trace is the end-to-end application flow from the entry point till the answer is produced, it is composed of smaller pieces called spans.
𝘾) Span is a smaller piece of the application flow that represents an atomic action like a function call or a database query. They can be sequential, or run in parallel.

ℹ️ As part of span we capture general metadata like start and end time, inputs and outputs of the span. On top of this metadata we track information specific to the GenAI system elements.

What might a trace look like for a naive RAG system?

𝟭. A query that has been submitted to the chat application.
𝟮. The query is embedded into a vector. 

✅ Additional metadata like input token count is persisted with the span so that we can estimate the cost of the procedure.

𝟯. ANN lookup performed against the Vector DB to retrieve the most relevant context. 

✅ Additional metadata about the query is persisted as part of the span together with the retrieved pieces of context and their relevance.

𝟰. A prompt is constructed from the system prompt and retrieved context.
𝟱. The prompt is passed to the LLM to construct the answer. 

✅ Additional metadata about input and output token count is captured together with the span so that we can estimate the cost of the procedure.

𝘞𝘩𝘺 𝘪𝘴 𝘵𝘳𝘢𝘤𝘪𝘯𝘨 𝘰𝘧 𝘎𝘦𝘯𝘈𝘐 𝘴𝘺𝘴𝘵𝘦𝘮𝘴 𝘪𝘮𝘱𝘰𝘳𝘵𝘢𝘯𝘵?

- These applications are usually complex chains, errors can happen in different steps of your application. E.g. Embedding of query is taking longer than expected or you have reached API limits of LLM provider.
- Cost for calling LLM APIs will be variable depending on the length of inputs and produced outputs. You would usually trace this information and analyze it to help forecast expenses.
- GenAI systems are non-deterministic and will deteriorate over time. They need to be evaluated on span level rather than input/output of the entire system so that you can tune each piece separately.
- …

Are you tracing your Agents? Let me know in the comments 👇

Don't get fooled, building a 𝗽𝗿𝗼𝗱𝘂𝗰𝘁𝗶𝗼𝗻 𝗴𝗿𝗮𝗱𝗲 𝗥𝗲𝘁𝗿𝗶𝗲𝘃𝗮𝗹 𝗔𝘂𝗴𝗺𝗲𝗻𝘁𝗲𝗱 𝗚𝗲𝗻𝗲𝗿𝗮𝘁𝗶𝗼𝗻 (𝗥𝗔𝗚) 𝗯𝗮𝘀𝗲𝗱 𝗔𝗜 𝘀𝘆𝘀𝘁𝗲𝗺 is a challenging task. Read until the end to understand why 👇

Here are some of the moving parts in the RAG based systems that you will need to take care of and continuously tune in order to achieve desired results:

𝗥𝗲𝘁𝗿𝗶𝗲𝘃𝗮𝗹:

𝘍 ) Chunking - how do you chunk the data that you will use for external context.

- Small, Large chunks.
- Sliding or tumbling window for chunking.
- Retrieve parent or linked chunks when searching or just use originally retrieved data.

𝘊 ) Choosing the embedding model to embed and query and external context to/from the latent space. Considering Contextual embeddings.

𝘋 ) Vector Database.

- Which Database to choose.
- Where to host.
- What metadata to store together with embeddings.
- Indexing strategy.

𝘌 ) Vector Search

- Choice of similarity measure.
- Choosing the query path - metadata first vs. ANN first.
- Hybrid search.

𝘎 ) Heuristics - business rules applied to your retrieval procedure.

- Time importance.
- Reranking.
- Duplicate context (diversity ranking).
- Source retrieval.
- Conditional document preprocessing.

𝗚𝗲𝗻𝗲𝗿𝗮𝘁𝗶𝗼𝗻:

𝘈 ) LLM - Choosing the right Large Language Model to power your application.

✅ It is becoming less of a headache the further we are into the LLM craze. The performance of available LLMs are converging, both open source and proprietary. The main choice nowadays is around using a proprietary model or self-hosting.

𝘉 ) Prompt Engineering - having context available for usage in your prompts does not free you from the hard work of engineering the prompts. You will still need to align the system to produce outputs that you desire and prevent jailbreak scenarios.

And let’s not forget the less popular part:

𝘏) Observing, Evaluating, Monitoring and Securing your application in production!

What other pieces of the system am I missing? Let me know in the comments 👇

If you want to start learning Kubernetes as an AI Engineer, look no further 👇

Some time ago I created a 2 part series on the main concepts you should know as a developer working with K8s. Examples and walkthroughs included - everything runs locally!

In the two articles you will learn:

➡️ Why Kubernetes for AI and Data Engineering.
➡️ General Kubernetes Resources for application configuration.
➡️ How to deploy applications.
➡️ Different ways to deploy applications and when to use which.
➡️ Visual explanation for all of the above.

Since I have resumed the newsletter, there will be many hands-on AI projects utilising Kubernetes for deployment. Be sure to sharpen your skills if you intend to follow! This resource is perfect for that.

Link to the first article: https://lnkd.in/ed2HfHjV

It is completely free 😉

Nationale-Nederlanden

Nationale-Nederlanden (NN) is a financial services company active in 11 countries. It provides retirement services, pensions, insurance, banking and investments to approximately 18 million customers.

I have been involved in the NNDAP project (NN Data Analytics Platform). NNDAP is a microservices based software product that aims to provide a data platform blueprint for every business unit of NN, maximising efficiency and minimising repetitive work.

With NNDAP, business units can easily provision a brand new data platform and deploy modules to it, such as a Data Extraction, Data Ingestion, Data Orchestration, Data Harmonization, Data Quality, oAuth, among others. All the modules can be flexibly configured via config-as-code, making sure deployment from development all the way to production is a breeze. Furthermore, NNDAP is proudly part of the 'Innersourcing' initiative (open-source within NN), allowing all engineers to contribute to the codebase of NNDAP.

Senior Data Platform Engineer

Rabobank

Set-up, development and maintenance of the Enterprise Data Lake for Rabobank in order to have an efficient, structured and reliable Enterprise Data Lake. This team is specialized in data governance topics, for the most part data quality and data catalogue integration. The Data Governance team provides self-service data governances capabilities and ensures integration with the data catalogue to enable that data is controlled, findable, understandable, accessible and meets regulations. We build solutions to comply to i.e., GDPR and define and develop solutions together with the team surrounding data quality checks.

Experience:

Strong knowledge about DevOps process (Git, CI/CD, YAML pipelines)
Strong coding skills (Python, FastAPI, Pytest)
Airflow
Good knowledge of Bash or PowerShell 
Good knowledge of Azure services ADLS gen2
Azure Data Factory (Pipelines & Data Flow)
Azure SQL server
Log Analytics Workspace
Security (Virtual Networks, Firewall and IAM)
Azure App Service
Azure Function
Azure Resource Manager Templates (ARM)
Bicep
Azure Active Directory
Power BI Dashboarding
Debugging Azure services from Activity Logs, Metrics, etc.
Experience with Data Modeling 
Data Warehouse Schema Design - Dimensional Modeling and Star Schema
Experience with Data Management topics like Data Governance, Data Privacy and Data Access
Scrum and DevOps methodologies

Data Platform Engineer (Data Governance)

FRISS

Helping to build solutions against fraud with: (ElasticSearch + FastAPI + Azure + Python + Graph/Network Analytics)

I basically work on 3 projects:

* Network Entity Lib: This is an Azure Artifact project, which provides a pip python package. It consists of a set of unified data models which are used internally in FRISS Products. It contains a representation of an Entity Graph (set of Entities and Edges), that we call FRISS Standard Model.

* Network ETL: Python application which transforms claims data in FRISS standard model (SDM) to a data format suitable to the network stack. It can be run in 2 modes:

- 1. Realtime: data in SDM format comes in via Kafka stream in real time.
- 2. Historical Load: A whole history of data in the form of blob documents from Azure Data Lake Storage is extracted, transformed and loaded into ElasticSearch database. This is either done when an existing Tenant will use the network stack for the first time or data needs to be transformed due to a change in the data model (Network Entity Lib)

* Network API: Python FastAPI application that serves the network viewer, network detection check, network claim history. It provides endpoints for searching, detection, expanding networks, storing tenant config and managing lists.

Senior Python Engineer (Fraud Analytics)

ABN AMRO Bank N.V.

Builds systems and models, which help the business to calculate, predict and prevent risks for internal and external customers.

* Gathering requirements from the business and refining them;
* Analysing the data;
* Modify  Credit Risk models developed by Data Science team and deploying them to Azure DevOps pipeline;
* Developing and adding new functionalities in Python which will enable the bank to better determine Financial Risk and create detailed automated Excel reports (internal and external reporting) with Python and openpyxl for Financial Auditing;
* Working in a DevOps way by taking over activities of other colleagues, like testing, implementing and production support;
* Support SAS team regarding Credit Risk models by providing required datasets and helping with functional tests.
* Help rewrite credit risk model features from Java code to Python
* Write and modify SQL queries on Oracle database according to business needs.

Python Software Engineer  (Risk Management)

Protolabs

I work building Supply Chain and Logistics features for 3D Hubs platform:

- Integration with Customer.io and Segment to send notification to clients when a certain event happens. Set Up a Task Queue with Celery and RabbitMQ
- Help transition Logistic process for European customers from Drop-ship model to Cross-Docking model.
- Interact with Auction allocation/bidding algorithm
- Create SQLAlchemy event listeners in order to help the Data Engineering team to gather useful information, mostly related to statuses timestamps.
- Surface backend features in the frontend: enable users to order 3D Hubs boxes and toggle CNC and 3D printing capabilities.
- Backfilling of data for new properties/columns created in database models.
- Enable Manufacturing capabilities, like CNC machining when new supplier is added to the platform (Technologies, Processes, Materials).
- Create API endpoints to be consumed in the fronted project.
- Architecting Stores with NgRx: build the entities that make the store, setting up the entity adapter, the actions, effects and the reducers for each entity.
- Add ability to view, edit, delete, create entities and properties in FlaskAdmin
- Work on New Order Manager Info component in order to show Shipping information to the user according to Order status.

Full Stack Engineer (Digital Manufacturing & Supply Chain)

Ericsson

I worked as a software developer contractor hired by Neueda, allocated in Ericsson site in Athlone, Ireland. I helped building backup and restore features in order to protect Ericsson’s clients data, which is related to the product Ericsson Management as a Service (ENMaaS). 

It basically evolves the following steps: encrypt, compress, decrypt, decompress and process the files while also having concerns about the performance, using Python multiprocessing library. Most of the commands sent to the servers were performed using the the Python subprocess module and bash commands. We used the rsync protocol, to transfer file volumes.

Regarding code quality, we used the Unittest framework and Mock library for Unit Tests and we used pylint and flake8 all configured in a tox automation file in order to check possible bugs, and style misconducts.

Regarding the versioning of the code we used Git and Gerrit Code Review and finally, for project management tracking, we used Jira aligned with Scrum methodology (daily meetings, retrospective and grooming).

Python Developer

Magazine Luiza

Grupo Mulheres do Brasil (Group Women of Brazil)" aims fostering female protagonism in several important areas for the development of Brazil as entrepreneurship, health and education. I helped them with the following:

- Redesign their  website: www.grupomulheresdobrasil.org.br.
- Manage product backlog.
- Set up payment system, so that the organization could receive donations.
- Participate in meetings with the stakeholders (Group Women of Brazil) in order to understand their needs and plan a strategy to achieve their goals.
- Organize database information for their intranet system and the deployed website.
- Perform migration operations and cleaning of the data.

Product Owner / Software Developer

Serasa Experian

I worked on the project Polis, I helped building a web API composed of functionalities to show the commute time among locations in the city of São Paulo, mobility index map, among others. The commute time was computed based on data collected and analyzed from SPTrans and Google Maps APIs. I used the following technologies: AWS, Python, Docker, Flask, Postgresql, Google Maps API. The frontend of the application used AngularJS 1, HTML5, CSS3 and Bootstrap and it basically showed the representation of the data processed in a meaningful way to the users.

Summer Job - Data Science/Engineering

Universidade de São Paulo

I worked with the subject Probability and Statistics. The Teaching Assistant Program (PAE) permits the active participation of Graduate Students in undergraduate teaching activities in courses offered by The University of Sao Paulo. The program offers the students the opportunity to train their teaching abilities in undergraduate classes under the supervision of experienced university teachers, preparing them for a future teaching career.

Teacher Assistant (Probability and Statistics)

Upwork

Worked with technical transcription and data entry for the company Pactera in order to help to improve the Skype voice recognition software.

Data Entry Specialist

* I want to be alive to see Python 3.14159. Wouldn't that be PieThon... or PiThon?

taplio

The Battle

Aurimas Griciūnas

👻 Nayana 👽 H.

Followers

Posts

Engagements

Likes

Performance ScoreAurimas Griciūnas's LinkedIn Performance

Ready to grow your LinkedIn presence?

taplio

The Battle

Aurimas Griciūnas

👻 Nayana 👽 H.

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);}Aurimas Griciūnas's LinkedIn Performance

Ready to grow your LinkedIn presence?

Performance ScoreAurimas Griciūnas's LinkedIn Performance